Transcript
Page 1: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

j ourna l homepage: www.e lsev ie r .com/ locate /pa laeo

Challenging the sensitivity limits of Paleomagnetism:Magnetostratigraphy of weakly magnetized Guadalupian–Lopingian(Permian) Limestone from Kyushu, Japan

Joseph L. Kirschvink a,b,⁎, Yukio Isozaki c, Hideotoshi Shibuya d, Yo-ichiro Otofuji e, Timothy D. Raub f,Isaac A. Hilburn a, Teruhisa Kasuya c, Masahiko Yokoyama e, Magali Bonifacie g

a Division of Geological & Planetary Sciences, California Institute of Technology 170-25, Pasadena, CA 91125, USAb Earth-Life Science Institute, Tokyo Institute of Technology, Meguro, Tokyo 152-8550, Japanc Dept. Earth Sci. Astron., Univ. Tokyo, Meguro, Tokyo 153-8902, Japand Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japane Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japanf Dept. of Earth Sciences, University of St Andrews, St Andrews, KY16 9AL Scotland UKg Institut de Physique du Globe de Paris, Laboratoire Géochimie des Isotopes Stables, IPGP, Bureau 515, rue Jussieu, 75238 Paris cedex 05, France

⁎ Corresponding author at: Division of Geological &Institute of Technology 170-25, Pasadena, CA 91125, USA

E-mail address: [email protected] (J.L. Kirschvin

http://dx.doi.org/10.1016/j.palaeo.2014.10.0370031-0182/© 2014 The Authors. Published by Elsevier B.V

a b s t r a c t

a r t i c l e i n f o

Article history:Received 5 July 2014Received in revised form 14 October 2014Accepted 28 October 2014Available online 4 November 2014

Keywords:PaleomagnetismLimestonePermianKiaman superchronIllawarra ReversalMass extinction

Despite their utility for bio- and chemostratigraphy, many carbonate platform sequences have been difficult toanalyze using paleomagnetic techniques due to their extraordinarily weak natural remanent magnetizations(NRMs). However, the physical processes of magnetization imply that stable NRMs can be preserved that aremany orders of magnitude below our present measurement abilities. Recent advances in reducing the noiselevel of superconductingmagnetometer systems, particularly the introduction of DC-SQUID sensors anddevelop-ment of a low-noise sample handling system using thin-walled quartz-glass vacuum tubes, have solved many ofthese instrumentation problems, increasing the effective sensitivity by a factor of nearly 50 over the previoustechniques of SQUID moment magnetometry.Here we report the successful isolation of a two-polarity characteristic remanent magnetization from Middle–Late Permian limestone formed in the atoll of a mid-oceanic paleo-seamount, now preserved in the Jurassic ac-cretionary complex in Japan, which had proved difficult to analyze in past studies. Paleothermometric indicatorsincluding Conodont Alteration Indices, carbonate petrology, and clumped isotope paleothermometry are consis-tent with peak burial temperatures close to 130 °C, consistent with rock magnetic indicators suggesting fine-grained magnetite and hematite holds the NRM. The magnetic polarity pattern is in broad agreement with pre-vious global magnetostratigraphic summaries from the interval of the Early–Middle Permian Kiaman ReversedSuperchron and the Permian–Triassic mixed interval, and ties the Tethyan–Panthalassan fusuline zones to it.Elevated levels of hematite associatedwith the positive δ13Ccarb of the Kamura event argue for a brief spike in en-vironmental oxygen. The results also place the paleo-seamount at a paleolatitude of ~12° S, in the middle of thePanthalassan Ocean, and imply a N/NW transport toward the Asianmargin of Pangea during Triassic and Jurassictimes, in accordance with the predicted trajectory from its tectono-sedimentary background. These develop-ments should expand the applicability of magnetostratigraphic techniques to many additional portions of theGeological time scale.

© 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license(http://creativecommons.org/licenses/by/3.0/).

1. Introduction

A fundamental goal of stratigraphy is to establish interbasin correla-tions with globally-isochronous time horizons, including the criticalstratotype horizons, and to correlate these to the magnetic reversal

Planetary Sciences, California. Tel.: +1 626 395 6136.k).

. This is an open access article under

patterns of the Geomagnetic Reversal Time Scale (GRTS), which shouldbe essentially isochronous on a global scale. Most of the globalstratotype sections and points (GSSPs) have been defined in fossilifer-ous, shallow marine carbonate platform sequences due to their provenability to record pristine biological and geochemical records of Earth his-tory. Unfortunately, it is well known that biostratigraphically-definedzone boundaries are often diachronous, and local oceanographic and geo-logical effects can influence geochemical proxies for chemostratigraphiccorrelation.

the CC BY license (http://creativecommons.org/licenses/by/3.0/).

Page 2: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

76 J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

Efforts to use paleomagnetic techniques for unravelingmagnetic po-larity patterns in many GSSPs, however, have often proven frustrating.Prominent examples include the original definition of the Silurian–Devonian stratotype at Klonk, in the Czech Republic, where themetamorphic grade appears to have been too high to retain primarypaleomagnetism (Ripperdan, 1990), and several Ordovician GSSPs onAnticosti Island, Canada (Ripperdan, 1990; Seguin and Petryk, 1986),wheremany of the pale carbonateswere found to be tooweaklymagne-tized to measure and demagnetize reliably using the conventional tech-niques of 25 years ago. From these studies it was simply not knownwhether the rocks had concentrations of magnetic minerals too smallto produce a measureable moment, or if diagenetic processes and hy-drocarbon migration had destroyed those that were there initially.

It has long been known that the physical processes of aligningmagnetic minerals during the formation of sedimentary rocks can pre-serve stable magnetic components that are many orders of magnitudebelow the measurement ability of the best superconducting rock(moment) magnetometers (Kirschvink, 1981), even those using the en-hanced sensitivity of DC-biased superconducting quantum interferencedevices (DC-SQUIDs) (Weiss et al., 2001). In fact, the situation is muchworse than this sensor noise limit, as the intrinsic magnetic moment ofmost of the sample holders used in paleomagnetic studies (backgroundnoise) is typically several orders of magnitude higher than the sensorlimit. DC-SQUID sensors in the commercially available 2G Enterprises™rock magnetometers typically have r.m.s. magnetic moment noise levelsof a few tenths of a pAm2 (a few ×10−10 emu), whereas most sampleholdersmeasure at a few tensof pAm2(10−7 emu), several orders ofmag-nitude higher. Recently, the introduction of acid-washed, thin-walledquartz-glass tubing for supporting samples with a vacuum has loweredthis noise greatly by minimizing the amount of extraneous matter in thesense region of the SQUID magnetometers (Kirschvink et al., 2008).Coupled with a computer-controlled pick-and-place sample changingsystem, this permits the large numbers of precise demagnetization exper-iments needed for magnetostratigraphic studies to be performed rapidly.

We chose the Middle–Late Permian limestone from Kamura inKyushu, Japan, to test the suitability of these new sample-measuringtechniques because previous rock magnetic work (Yokoyama et al.,2007) demonstrated that these rocks contained fine-grained magnetiteand hematite, but in concentrations making the NRM difficult toanalyze. Most of the Jurassic accretionary complex in SW Japan that con-tains exotic Permian limestone blocks suffered at most lower greenschistfacies metamorphism around 140 Ma (Isozaki et al., 1990), usuallycharacterized by the mineral paragenesis of pumpellyite–actinolite, ex-cept some locally baked domains in close contact with the Cretaceous–Paleogene granitic intrusions. Although the Kamura area in central Kyu-shu is located near an active volcanic region, there are ample indicationsfrom the local geology that these rocks were never affected significantlyby thermochemical alteration. The studied interval straddles theMiddle–Late Permian boundary (Isozaki and Ota, 2001; Ota andIsozaki, 2006) and thus records the end-Guadalupian mass extinctionevent (Jin et al., 1994; Stanley and Yang, 1994) and possibly the top ofthe Kiaman Reversed Superchron, known as the ‘Illawarra Reversal’(Cottrell et al., 2008; Courtillot and Olson, 2007; Gialanella et al.,1997; Gradstein et al., 2012; Irving, 1964; Isozaki, 2009; Opdyke et al.,2000), offering the possibility of enhancing the correlation betweenthe two time scales. In particular, the first appearance of a solid Normalinterval is critical in identifying the Illawarra Reversal that is expectedwithin Wordian (Middle Guadalupian) time.

Application of these new techniques to the Kamura limestone revealsthe presence of a stable, 2-polarity characteristic NRM that is broadlyconsistent with past studies of the geomagnetic polarity chronology forlate Permian Time, including the top of the Kiaman Superchron. Thecharacteristic direction, and the match to the reversal chronology, indi-cates that the Kamura atoll was located at about 12° South latitude inthe Panthalassic Ocean. A minimum of 3000 km of N/NW transportwould have been required for it to dock against the Eastern margin

of Pangea during Jurassic time. The present result also has profound im-plications to the bio- and chemostratigraphic correlation between themid-superoceanic paleo-atoll limestone and continental shelf carbonatesaround Pangea.

2. Geological setting

2.1. Tectono-sedimentary background

The Permian and Triassic limestone at Kamura (Takachiho town,Miyazaki prefecture; Fig. 1) in Kyushu forms a part of an ancient mid-oceanic atoll complex primarily developed on a mid-oceanic paleo-seamount (Isozaki, 2014; Isozaki and Ota, 2001; Kasuya et al., 2012;Ota and Isozaki, 2006; Sano and Nakashima, 1997). This limestone oc-curs as a several kilometer-long, lensoid allochthonous block withinthe Middle–Upper Jurassic disorganized mudstone/sandstone of theJurassic accretionary complex in the Chichibu belt, southwest Japan,with remarkably little internal deformation (Fig. 2). The orientation ofthe late Paleozoic to early Mesozoic subduction zone beneath theAsian blocks (Isozaki, 1997a,b) implies that the seamount originatedto the east (Pacific side) with respect to Asia, i.e., somewhere inthe superocean Panthalassa, and accreted to the Asian margin in theJurassic, approximately 100 million years later. The limestone blocks inthe Kamura area retain parts of the primary mid-oceanic stratigraphy(ca. 135m in thickness), and range fromWordian (middle Guadalupian)to Norian (Upper Triassic) time with several sedimentary breaks in theTriassic part (Isozaki, 2014; Kambe, 1963; Kanmera and Nakazawa,1973; Kasuya et al., 2012; Koike, 1996; Ota and Isozaki, 2006).

2.2. Lithostratigraphy and paleoenvironments

The Iwato Formation consists of ca. 100 m-thick, dark gray to blackbioclastic limestone. Bioclasts include fragments of bivalves, calcareousalgae, crinoids, fusulines and other small foraminifera, indicatingGuadalupian age. The lower part of the Iwato Formation compriseswackestone with a black, organic-rich matrix and yields abundantlarge bivalves (Family Alatoconchidae) and large-tested fusulines(e.g., Neoschwagerina, Yabeina, Lepidolina). The upper part comprisespeloidal wackestone. Black organic matter probably of microbial originis concentrated in peloids.Megafossils are absent in this interval, exceptfor very rare rugose coral (Liangshanophyllum) from the uppermostpart. All black limestones are free from dolomitization.

The overlying Mitai Formation consists of nearly 40 m-thick, lightgray bioclastic dolomitic limestone. Bioclasts are derived from calcareousalgae, crinoids, ostracodes, gastropods, bivalves, crinoids, brachiopods,coral, fusulines and small foraminifera of the Tethyan affinity, and indi-cate Lopingian age. The limestones are mostly grainstone/wackestonewith lesser amounts of lime–mudstone that are fossiliferous, mostlymassive, partly including 1 cm-thick, continuous to discontinuousbandswith concentrations of peloids and algae. Crystals of secondary do-lomite are generally concentrated around bioclasts (and avoided in thepaleomagnetic sampling). The lowermost 1 m-thick bed is characterizedby white bands containing abundant dolomitized dasycladacean algae.

All the Tethyan fusuline assemblages and associated fossils from theIwato andMitai formations indicate that the seamount was located in alow-latitude domain in the superocean Panthalassa under a tropicalclimate (Isozaki, 2006, 2014; Isozaki and Aljinovic, 2009; Kasuya et al.,2012).

2.3. Bio- and chemostratigraphy

Conodonts, the index fossils with the highest resolution for thePermian, have unfortunately not been found in the Permian Iwato andMitai formations where our paleomagnetic samples come from, as thesedimentary facies was too shallow to host conodont animals. Fusulinesare the most abundant among fossils, and they provide a basis for

Page 3: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

Pan

gea

Kamura

A

B C

Mino-Tanba belt

Chichibu belt

130˚

E 35˚N

135˚

E

Jurassic (latest Triassic to earliest Cretaceous) accretionary complex

300 km

Kyushu

shikoku

Honshu

paleo-atoll

paleo-seamount

accretionary complex

oceanic plate

Fig. 1. (A) Index map of southwest Japan, showing the distribution of the Jurassic accretionary complex with accreted paleo-atoll limestones of the Chichibu belt, extending from centralKyushu, onto Shikoku and western Honshu. (B) A simplified cartoon of the paleo-ridge-arc transect interpreted to have been operating from Late Permian through mid-Jurassic time.(C) Paleogeography of the Permian world with the probable location of the Kamura seamount (filled circle). Adapted from Kasuya et al. (2012).

Sec. 8 KAM

Sec. 1

Sec. 3 SARAW

Sec. 2 SARA

Sec. 4Sec. 5

Sec. 6 SHIO-S

Sec. 7

Sec. 9 SHIO-W

Sec. 10 HJPM

Sec. 11Sec. 12

Shioinouso

Koseri

Saraito

32°45′N

131°20′E 131°21′E

Mitai Formation (Lopingian)

Iwato Fomation (Guadalupian)

Kamura Formation (Triassic)

Sections Sampled for Paleomagnetics

500 m0

Fig. 2. Geological sketch map of the Kamura area in central Kyushu, showing the sections analyzed for paleomagnetics and the distribution of the Permo–Triassic paleo-atoll limestonesections. This figure is adapted from Kasuya et al. (2012). Beds within the exposure are nearly vertical to slightly overturned, striking generally ENE/WSW, and younging to the North.

77J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

Page 4: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

78 J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

subdividing the Iwato Formation into the following 5 biostratigraphicunits; i.e., the Neoschwagerina craticulifera Zone, Neoschwagerinamargaritae Zone, Yabeina Zone, Lepidolina Zone, and a barren interval,in ascending order (Kasuya et al., 2012; Ota and Isozaki, 2006). Strati-graphic positions of the zone boundaries are constrained by fossil occur-rence to be better than 1 m, except for the N. craticulifera/N. margaritae,boundarywhich is estimated to be ~±5m. TheN. craticulifera Zone andN. margaritae Zone are correlated with the Wordian of Texas and withtheMurgabian and lowerMidian in Transcaucasia, whereas the YabeinaZone and Lepidolina Zone, andmost of the barren interval are correlatedwith the Capitanian (Upper Guadalupian) of Texas and with the mid-dle–upper Midian in Transcaucasia (e.g., (Wardlaw et al., 2004)). Thecarbon isotope stratigraphy of the Lepidolina Zone and barren interval(Isozaki, 1997a,b) documented the occurrence of high positive plateauin δ13Ccarb values (N5‰) that can be correlated with the Capitanianrocks in North America and the European Tethys (e.g. (Isozaki et al.,2011)). In addition, the uniquely low Sr/Sr ratios (b0.7070) detectedin the same interval also support the age assignment of the Capitanian(Kani et al., 2008; Kani et al., 2013).

The overlying Mitai Formation belongs to the Codonofusiella–Reichelina Zone that corresponds to the Wuchiapingian (LowerLopingian) in South China. The carbon isotope stratigraphy alsoconfirmed that the gradual positive excursion in the lowermostCodonofusiella–Reichelina Zone (Isozaki et al., 2007b) correspondsto the same signal detected in the Clarkina dukoensis Zone (lowerWuchiapingian) at Penglaitan, the GSSP (Global Stratotype Section andPoint) of G–L boundary in South China (Chen et al., 2011; Wang et al.,2004).

2.4. Geochemistry

In terms of total organic content (TOC), the black limestone of theIwato Formation varies between 0.13 and 0.77 wt.%, whereas the lightgray dolomitic limestone of the Mitai Formation varies from 0.0044 to0.0067 wt.% (Isozaki et al., 2007b), values which are typical of oceanicatoll environments. However, the carbonates in the Wordian to lowerCapitanian interval at Kamura record a major excursion in the valueof δ13Ccarb, increasing from a stable value of ~+4.5‰ and reaching amaximum of ~+7.0‰ within the Yabeina Zone (fusuline) of the early–middle Capitanian. Thus the total duration of the Kamura event is esti-mated ca. 3–4 million years, occupying the majority of Capitaniantime (Isozaki et al., 2007a).

In addition to the positive excursion in carbon isotopes, the Kamuralimestones also preserve a record of the Paleozoic minimum in marine87Sr/86Sr ratio, which reaches its nadir (0.706914 ± 0.000012) in lateCapitanian time, essentially coincident with the Kamura isotope excur-sion (Kani et al., 2008).

2.5. Constraints on burial metamorphism

Although the Kamura limestone unit we have studied is a coherentblock that crops out over a large area (~10 km long, ~150 meter thick,with bedding nearly vertical, and remarkably similar structural atti-tudes), it occurs as large exotic blocks within the younger Jurassic sand-stone/mudstone along the length of the accretionary complex shown inFig. 1 (Ota and Isozaki, 2006). It has long been known that limestone‘knockers’ scraped off of subducting slabs often escape significant burialmetamorphism under these conditions, as exemplified by the classicpaleomagnetic study of the Laytonville Limestone in the Franciscancomplex of central California (Alvarez et al., 1980). The sequence inKyushu, however, has been intruded subsequently by felsic intrusionsof arc affinity, and is in the vicinity of Mt. Aso, a major modern eruptivecenter on the island of Kyushu that is capable of producing local and/orregional thermochemical alteration. Paleomagnetic studies of Mesozoicrocks in both Kyushu and central Honshu sometimes show a complexpattern of remagnetization, indicating both thermal and chemical

events have operated in some areas (Abrajevitch et al., 2011;Abrajevitch and Kodama, 2009; Ando et al., 2001; Kodama and Takeda,2002; Oda and Suzuki, 2000; Shibuya and Sasajima, 1986; Uno et al.,2012).

Despite this, three observations suggest that the rocks that we arestudying have not been altered to levels capable of erasing completelytheir initial remanent magnetizations. First, the shells of the giant bi-valves appear remarkably pristine. They are built with a typical mollus-can double-layered structure, where the external layer is composed ofparallel-aligned prismatic calcite with c-axis perpendicular to the shellsurface, whereas the inner layer is a micro-crystalline mosaic (Isozaki,2006; Isozaki and Aljinovic, 2009). This textural difference arguesagainst thermochemical alteration high enough to trigger the pervasiverecrystallization of the carbonate, and certainly well below any carbon-ate decomposition temperature.

Second, conodonts in the overlying Triassic portions of the block areoften preserved with Conodont Alteration Indices (CAI) as low as 2 onthe scale of Epstein et al. (1977), implying only mild heating (60°–140 °C) on geological time scales, but peak heating of up to ~500 °Con the time scale of less than an hour (which is implausible in thegeological setting of these rocks), (Ripperdan, 1990). Hence, a blanket,regional heating capable of complete remagnetization of the sedimentscan be ruled out, particularly as the most likely magnetic carriers(magnetite and hematite) are stable to much higher temperatures.

Unfortunately, conodonts have not been found in the Permian por-tion of the limestone in Kamura where our paleomagnetic samplescome from, so their CAI values cannot be used to test for more signifi-cant thermochemical alteration from local intrusions or volcanic vents.However, field observations of localities where intrusions do impingeon the limestone reveal intense and visually obvious bleaching of therock as the kerogens are destroyed. Our paleomagnetic sampling hasavoided these areas. Further constraints from clumped isotope analysesare presented in Section 5.3 below.

3. Studied sections and rock samples

Westudied 3major sections in the Kamura area; i.e., Sections 2, 8, and10 (Figs. 2, 3) with additional samples from 3 other sections (Section 4, 7,and 9; see Kausya et al. (2012) for details). Section 2, which crops outat the southeast of Saraito village (32° 45′12″ N, 131° 20′55″ E), contains57 m of black limestone that belongs to the N. craticulifera Zone (Isozaki,2006; Isozaki et al., 2007a; Kambe, 1963). Owing to the outcrop condi-tion, we only sampled from the lower half of the section. Section 8 atthe west of Shioinouso village (32° 44′58″ N, 131° 20′02″ E) contains35 m of similar limestone that contains the upper Iwato Formationand the lower Mitai Formation, and spans the Guadalupian–Lopingianboundary (G–LB) as originally described by Ota and Isozaki (2006).We collected samples from the Capitanian Lepidolina Zone, overlyingbarren interval, and the Wuchiapingian Codonofusiella–Reichelina Zone.Section 10 at Hijirikawa consists of 9 m of thick black limestone of theIwato Formation that belongs to the Yabeina Zone (Isozaki et al.,2007b). We collected samples from the entire section. As to the other 3sections, Section 4 belongs to the Neoschwagerina margaritae Zone,whereas Sections 7 and 9 belong to the Yabeina Zone (Kasuya et al.,2012). Although there is no single continuous section that covers the en-tire Guadalupian stratigraphy in this area, wemanaged to collect samplesfrom all fusuline zones ranging from theWordian toWuchiapingian, andin particular, at the continuous Section 8 that preserves a 35 m-thicklimestone across the G–LB.

For the magnetostratigraphic study, we collected a total of 117oriented core samples (2.5 cmdiameter cylinders) of fine-grained lime-stone (lime mudstone composed of pure carbonates with scarce terrig-enous components) from the above-described 3 sections in the Kamuraarea (Sections 2, 8 and 10), using standardmagnetic and solar compasstechniques. In addition, eleven hand samples fromSections 4, 7 and 9, inlocations difficult to use the portable drills, were oriented in-situ using a

Page 5: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

0 m

100

50

Assemblage Zone 1(Neoschwagerina craticulifera Z.)

Assemblage Zone 2(Neoschwagerina margaritae Z.)

Assemblage Zone 3(Yabeina Z.)

Assemblage Zone 4(Lepidolina Z.)

barren interval

Codonofusiella-Reichelina Zone

Cap

itani

anW

uchi

apin

g.W

ordi

an

?

transitional interval between Zone 3 and Zone 4

Nc

NmNm

N+G

Y

Nc

L

LL

Y+LY

L

G

YY+GGGY+N

Nc

YY+Nm

Nm

Nm

Nm

G

YYGY+G+N

L+G

Y

Nm

N

Nc

NmY

NN

Y

Nc

Nm mar

Nc

Y :

G : L :

Iwat

o F

mM

itai F

m fusulineAlatoconchidae

Sec. 2

Sec. 11

Sec. 10

Sec. 8Sec. 7

Sec. 5

Sec. 4

Sec. 3

Fusulines

Sec. 9

N :

Fig. 3. Biostratigraphic correlation based on Fusuline zones for the localities sampled in the Kamura area. The section numbers correspond to those shown in Fig. 2. Assemblage zones 1–4correspond to the standard Tethyan nomenclature.

79J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

custom magnetic and sun-compass system and removed for samplepreparation in the laboratory. Forty additional block samples for inten-sive rock magnetic investigation were collected from Section 8 in theCapitanian and Wuchiapingian parts.

4. Materials & methods

4.1. Samples for paleomagnetic measurements

In the laboratory, the core sampleswere trimmed into asmany 1-cmhigh flat-ended discs as possible using non-magnetic saw blades,and the innermost specimens from each core (furthest from surfaceweathering) were selected for initial measurement processes. Similarcores were drilled from the eleven oriented hand samples using adrill press setup, with between 1 and 3 cores per sample, and werereoriented and labeled using standard techniques. Oriented end chipsfrom a representative batch of the cores from each locality were setaside for intensive rock magnetic analyses.

Due to the extremely weak magnetizations reported for these rocksby Yokoyama et al. (2007), we adopted moderately stringent clean-labhandling processes to minimize the possibility of stray ferromagneticmaterials becoming attached to the rock surfaces during the measure-ment and demagnetization procedures. First, after cutting and trimmingthe samples with the diamond-impregnated drills and saw blades, andenhancing the fiducial orientation line with a diamond scribe, theywere dipped vigorously but briefly (for less than 1 s) in concentrated,reagent-grade (12 N) HCl using plastic tongs, and then rinsed quicklyin a large volume of deionized, fresh water. This acid treatment causedan immediate reaction with the surface carbonate on the samples,noticeably darkened the rock surface, and often left an oily, blackresidue floating on the surface of the cleaning liquids. Care had to betaken not to lose either the fiducial marks or the sample identificationlabels during this process. Next, specimens were labeled with a non-magnetic white, thermally-resistant ink, and were moved into themagnetically-shielded environment (b200 nT) surrounding the SQUIDmagnetometerwork area for at least several days prior tomeasurement.The specimens, and all surfaces that touched them,were dusted off withhigh-pressure air thatwas passed through a 0.2 micron particulate filterprior to every measurement. We used disposable dust-free plastic

gloves for handling all of the samples once the demagnetization proce-dures were started.

4.2. Paleomagnetism

All remanence measurements for this study were conducted onthe Eugene M. Shoemaker Memorial Magnetometer, which is a 2G™Enterprises model 755 superconducting rock magnetometer with3-axis DC SQUIDS housed in a double-layer, mu-metal shielded roomat Caltech. Sham measurements of the baseline noise on the system(runwith everything operating except the quartz-glass vacuum sampleholder) yield a repeatable threshold moment sensitivity of a few by10−10 emu (~10−13 Am2). Inclusion of the quartz-glass sample tubes(19 mm diameter with 1 mm thick walls) usually raised the baselineholder moment up by a factor of 10 to 100 above this, which is far toohigh for successfully measuring the set of Kamura limestone sampleswe were investigating. To knock down the moment of the sampleholders, we first washed the tubes thoroughly with laboratory glasscleaners, and then soaked one end in reagent-grade, 12NHCl for severaldays to dissolve any ferric contaminants on and/or within the glassthat could be reached by the acid. For about a quarter of the glasstubes, this treatment resulted in holder moments below about10−9 emu (10−12 Am2 = 1 pAm2). On occasion they could becomemagnetically ‘invisible’, consistently reaching down to the backgroundnoise level of the DC SQUID sensors. In addition, we implemented aroutine in the RAPID software that allows the empty sample holder tobe cleaned with maximum strength alternating-field (AF) demagneti-zation after every set of nine measurements.

To place this in context, we note that our standard specimen size(cylinders ~2.54 cm in diameter, by ~1 cm high) is about 40% smallerthan the standard sample size traditionally used in most paleomagneticlaboratories (e.g., (Butler, 1992)). We did this because the smallerspecimen volume allows twice as many samples to be put into themagnetically-shielded furnaces for thermal demagnetization (up to~80 specimens per load), doubling the measurement rate. It also re-duced the problem of jamming on early versions of the automatic sam-ple changer used by the RAPID consortium (Kirschvink et al., 2008).Note that we prefer to discuss sensitivity in terms of magnetic moment,rather than magnetization (moment per volume), simply because the

Page 6: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

80 J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

raw data from the 2G™ magnetometers are, by the physics of their op-erating principles, sensitive to the total magnetic moment of a speci-men. SQUID magnetometers monitor the differential change in electriccurrent flowing in a pair of Helmholtz-geometry superconducting pick-up rings, and that is directly proportional to the total magnetic fluxchange across the loops, which is in turn proportional to the magneticmoment of a sample inserted into the center of the sense region(Fuller et al., 1985). The shape of the sample is simply not important(flat discs, cubes, or irregular objects are all OK), as long as the materialbeing measured is brought within the uniform region of the sensor(Kirschvink, 1992). This is different from older methods of measure-ments such as the astatic, induction-spinning, and fluxgate-ring sensorswhere the shape of individual samples was critically important. Com-paring the specimenmoment to that of the sample holder is also confus-ing if one is measured in terms of magnetization, rather than in units ofmoment.

After initial measurement of the natural remanent magnetizations(NRMs), sampleswere subjected to two ormore low-temperature ther-mal cycles in liquid nitrogen to help remove any viscous componentsthat might have been carried by multi-domain magnetite. This wasfollowed by a series of low-intensity progressive 3-axis alternating-field (AF) demagnetization experiments evenly spaced up to about7 mT, again to help remove any soft magnetic components that mighthave been acquired by exposure to moderate magnetic fields fromAirport X-ray machines or during the sample preparation. This wasfollowed by progressive thermal demagnetization under a gentle flowof N2 gas in a magnetically-shielded oven (b25 nT residual field),starting at 75 °C and incrementing in steps ranging from 10 to 25 °C,until the magnetization vectors were too weak to measure or the spec-imens displayed unstable behavior. We have found that heating in theN2 gas helps reduce the oxidation of ferrimagnetic minerals, reducingthe problem of the acquisition of spurious components upon mid-level heating steps. If the first specimen from a core displayed erraticor unstable behavior before isolating or revealing the characteristiccomponent, an additional specimen from the sample was run througha similar demagnetization series. To check for chemical changes duringthermal demagnetization, we measured the bulk susceptibility aftereach demagnetization step for each sample using a Bartington™ MS-2susceptibility bridge that was built into the sample changing systemas described elsewhere (Kirschvink et al., 2008). Between 15 and 30discrete demagnetization experiments were performed per specimen(average 24), on over 320 specimens, for an overall data set of over7800 complete measurements of NRM. Principle magnetic componentswere determined using the techniques of Kirschvink (1980), with theparameter for the Maximum Angular Deviation (MAD) set at 15°. Dueto theweaknature of theNRMs, components thatwere trending towardthe origin, or formed ‘stable endpoint clusters’ before the onset of unsta-ble behavior, were fit by forcing the least-squares line through the ori-gin (anchoring).

4.3. Rock magnetism

As the Kamura limestone possesses an unusually weak, but stable,NRM, we conducted a series of rock magnetic investigations to unravelthe nature of the magnetic phases present. At the KAM locality, speci-mens cut from the 40 block samples in the Kobe University collectionfrom across the G/L boundary were subjected to isothermal remanence(IRM) acquisition experiments to access for the presence of ferromag-netic and antiferromagnetic mineral phases. For a sub-set of thesespecimens we followed the general procedure of Lowrie (1990) foridentifying themagneticminerals present based on the distinctive coer-civity and blocking temperature characteristics of hematite, magnetite,and phyrrhotite, respectively. Hematite has both the highest Néeltemperature (670°) and coercivity (N1 T) of these minerals, the Néeltemperatures of the pyrrhotite mineral family are at or below ~325 °Cwith intermediate coercivities between ~0.4 and 1 T (Dekkers, 1989),

whereas magnetite has a Néel temperature of 580 °C and coercivitiesbelow 0.3 T. The progressive thermal demagnetization characteristicsof these components are thus diagnostic of these minerals, and can beseparated by giving a set of orthogonal isothermal remanentmagnetiza-tions (IRMs) of progressively lower value (Lowrie, 1990). We thereforeconducted thermal demagnetization of orthogonal IRMs given sequen-tially at 2.7 T, 400mT, and 120mT, at progressively higher peak temper-atures up to 680 °C.

To further characterize the NRM, a series of specimens from thesame section (8) were subjected to the suite of non-destructive rockmagnetic experiments described by Kirschvink et al. (2008) at Caltech,startingwith the3-axis Alternating-field (AF) progressivedemagnetiza-tion of the NRM in peak fields up to 80 mT. This was then followed bythe anhysteretic remanent magnetization (ARM) version of theLowrie–Fuller test for single-domain behavior (Johnson et al., 1975),which involved progressive acquisition of an ARM in peak alternating-fields (AF) of 100 mT, with variable DC biasing fields of 0–1 mT, follow-ed by progressive AF demagnetization of the maximum ARM. In thistest, the ARM substitutes for a weak-field TRM for single-domainand pseudo-single-domain particles, and provides a simple methodfor evaluating the relative contribution of these to the less geologicallystable multi-domain fraction. Samples were then given an IRM pulsein a peak field of 100 mT, followed by progressive AF demagnetizationexperiments (providing the ARM Lowrie-fuller tests for the fractionof the NRM with coercivities below 100 mT, e.g., (Johnson et al., 1975)).Finally, the sampleswere subjected to progressive IRMacquisition exper-iments up to 350mT, again followed by progressive AF demagnetization.

4.4. Clumped isotopes

Carbonate clumped isotope thermometry (Δ47) is based on thetemperature-dependent preference of rare isotopes 13C and 18O tobond with each other within a carbonate lattice. Due to the strongerchemical bond energy, the fraction of CO2 molecules that contain both13C and 18O atoms decreases as the temperature of crystallizationincreases; recently-developed techniques and calibration allow this‘clumpiness’ to bemeasured easily. This thermometry allows determin-ing independently the temperature at which carbonate crystallized andthe δ18O of the mineralizing fluid (e.g., (Ghosh et al., 2006)), and hasmostly been used for paleoclimatological reconstructions (e.g., (Eiler,2011; Ghosh et al., 2006) and references therein). However, whencarbonates experience high temperatures from burial or contact meta-morphism, solid-state diffusion of C and O within the mineral latticemight occur, leading to some reordering of 13C–18O bonds on both theprograde and retrograde paths during a geological heating event(Bonifacie et al., 2011; Passey and Henkes, 2012). This has been pro-posed based on studies of metamorphic marbles that generally showapparent equilibrium temperatures of about ~190 °C (with Δ47 averag-ing 0.352‰; (Bonifacie et al., 2011; Ghosh et al., 2006)) despite peakmetamorphic temperatures far above 500 °C. This apparent tempera-ture for marble is thought to represent the “blocking” temperaturewith respect to diffusional resetting of the carbonate clumped isotopethermometer — that is the temperature at which the isotopes stopexchanging by solid-state diffusion during gradual cooling (Bonifacieet al., 2011; Ghosh et al., 2006). In deeply buried carbonates thisphenomenon could therefore potentially challenge, under somecircumstances, the reconstruction for both the original temperatureof crystallization and the δ18O value of the mineralizing fluid (be-cause the original distribution of 13C–18O bonds could be progres-sively lost over heating while the oxygen isotopic composition ofthe carbonate could still be unaffected if dissolution/recrystallizationdid not happen), but it could also provide a way to reconstruct burial/heating path conditions (e.g. Bonifacie et al., 2011; Eiler, 2011; Passeyand Henkes, 2012).

For this work, CO2 was released from four samples of limestone,three from the Capitanian Iwato Fm (KAM7-7-2 Shell, KAM7-7-2

Page 7: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

Table 1Clumped isotope results from samples of the Kamura LimestoneCaption = Δ47 are reported in‰ relative to theGhosh scale, that is compared to a stochas-tic distribution of 13C–18O bonds in CO2 that have been heated to 1000 °C (Ghosh et al.(2006)). External precisions on replicate measurements of the same powder are typically±0.11‰ for both δ18O and δ13C and 0.015‰ on Δ47 measurements (representing here onaverage ±13–20 °C on temperature estimates when one consider the current uncer-tainties on the equation relating Δ47 to temperatures for high temperature materials).

Sample n δ13C δ18O Δ47 Temperature

ID (‰, PDB) (‰, PDB) (‰, Ghosh scale) (°C)

KAM7-7-2 Shell 2 4.52 −10.57 0.388 140KAM7-7-2 Mtx 3 3.79 −10.02 0.414 120KAM124.2 2 3.35 −7.87 0.406 125KAM38.2 3 5.60 −9.73 0.391 138Standard CarraraMarble

10 2.30 −1.72 0.350 190

81J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

Matrix, and KAM38.2 in Table 1) and one from theWuchiapingianMitaiFm (KAM124.2), using analytical procedures as well as standardizationand normalization procedures similar to Ghosh et al. (2006) (i.e., basedon CO2 gases equilibrated at 1000 °C and one carbonate marble refer-ence material). Purified CO2 was then analyzed at Caltech for its stableisotopic compositions (δ18O, δ13C and Δ47) on a Finnigan MAT 253 gassource mass spectrometer, configured to collect masses 44 through 49.Measured values of Δ47 (reported in‰ compared to a stochastic distri-bution of 13C–18O bonds in CO2 that have been heated at 1000 °C,defined in Ghosh et al. (2006)) were used to estimate carbonate appar-ent temperatures, using the empirically derived polynomial determinedusing high temperature experimental calcites (Guo et al., 2009), hydro-thermal dolomite (Bonifacie et al., 2011) and inorganic synthetic cal-cites (Ghosh et al., 2006) all generated into the same intra laboratoryframe than the samples investigated here. All samples were analyzedat least twice using sub-fractions of the same powder, to account forheterogeneity.

5. Results

5.1. Composite IRM analysis (Kobe University)

IRM acquisition and thermal demagnetization of composite-IRMexperiments shown in supplemental Figs. S1–S5 revealed that magnetiteexists ubiquitously through thewhole sampled section in concentrationsthat are almost constant in both the Iwato and the Mitai Formations,except for around the G/L boundary interval. Hematite, however, onlyexists in the Iwato Formation as the predominant mineral (in terms ofbulk volume).

The rock magnetic transition between the Iwato and the MitaiFormations does notmatch the lithologic boundary, as this change is lo-cated a few meters below it. The high proportion of hematite to othermagnetic minerals in the Iwato Formation is correlated positively withthe Kamura carbon isotope excursion in the same section. Diageneticprocesses such as consumption of hematite in the Mitai Formation orformation of hematite in the Iwato Formation are unlikely to explainthe predominance of hematite in the Iwato Formation because of thepresence of a constant amount of magnetite in both of the Iwato andMitai Formations. An interpretation of transport from terrestrial originof hematite is also discarded because of lithologic features of these for-mations that argue for deposition in an isolated island atoll carbonateplatform. Consequently, the rock magnetic properties of the Iwato andthe Mitai Formations are most likely depositional features, and thepresence of hematite associated with the positive rise in stable carbonisotope ratios probably reflects a change in the paleoenvironmentin the superocean Panthalassa, consistent with a rise in environmentaloxygen associated with increased carbon burial.

5.2. Coercivity spectral analysis (Caltech RAPID system)

Results from the coercivity analysis generally support the conclusionfrom the composite IRM analysis, as shown in Fig. 4. The IRM/ARMcoer-civity spectral analysis of Cisowski (1981) (top row of Fig. 4) shows abroad variability in the Iwato formation, with medium destructivefield values spreading between 40 and 60 mT, but with some samples(e.g., KAM 9 and 13) barely starting to approach saturation at peakpulse fields of up to 300mT, indicating the presence of a high-coercivityantiferromagnetic phase like hematite. All samples from the Mitaiformation, however, have much more reproducible coercivity spectra,with medium destructive fields in the 30–40 mT range, and clear ap-proach to saturation by 300 mT suggesting a much smaller antiferro-magnetic component. The two formations also differ considerablyin their interparticle interaction characteristics as determined by ARMacquisition in the coercivity band b100 mT (second row of Fig. 4).Data from the Iwato formation plot halfway between the chiton toothand magnetotactic bacterial reference curves, indicating the presenceof a mixture of interacting and non-interacting particles. In contrast,the Mitai samples follow a trajectory indicating significantly less inter-particle interaction, much more like that of partially collapsed bacterialmagnetosomes, as calibrated by Kobayashi et al. (2006). Magneticparticles in both formations are dominated by single-domain (SD) orpseudo-single-domain states as indicated by the ARM version of theLowrie–Fuller test (Johnson et al., 1975) as shown in the 3rd row inFig. 4. On all samples, the curve for the progressive AF demagnetizationof the ARM lies on top of that for the AF of the IRM, indicating domina-tion by SDparticles. However, the interparticle interaction effects (givenby the relative separation between the two curves) are clearly strongerfor the Iwato formation than for the Mitai, suggesting that somethinghas acted to permit more magnetic clumping.

Fuller et al.'s (1988) test of NRMorigin (bottom row in Fig. 4), whichcompares the intensity of the NRM remaining during AF demagnetiza-tion with that of the IRM, strongly supports the interpretation that theNRM signal is a depositional or post-depositional remanent magnetiza-tion (DRM or pDRM), rather than a CRM or TRM. This is because thevalues are nearly 3 orders of magnitude less than the correspondingIRM levels, and nearly 2 orders below the ARM values.

A scan using the Caltech ultra-high resolution SQUID magneticmicroscope (Weiss et al., 2001) on a polished surface of a sample fromthe upper portion of the Iwato Formation (KAM-109, Fig. S6) revealsthat the magnetization is diffusely located through the material, ratherthan being localized in discrete clumps or sedimentary grains.

5.3. Clumped isotopes

Table 1 shows results from the four samplesmeasured for this study.Measured carbon isotope values (δ13Ccarb) for limestone samples fromthe Kamura event in Kyushu (Isozaki et al., 2007a) are typically in therange of +3 to as high as +6‰, and oxygen (δ18Ocarb) is in the rangefrom −11 to −8‰; the values obtained here are in reasonably goodagreement. Furthermore, Δ47 values of the four investigated samplesare remarkably similar (average of 0.400‰ ± 0.012‰ standard devia-tion) within uncertainties on replicate Δ47 measurements (typically±0.015‰ with respect to standards). When converted into tempera-ture estimates, these results converge toward apparent metamorphictemperature of ~130 °C, suggesting that the actual peak temperaturewas probably close to this. On the other hand, it is worth noting thatat elevated temperatures and particularly where apparent Δ47 valuesapproach those seen in marbles, it is very difficult to easily rule outthe potential contribution of solid-state diffusion that might have onlypartly reset bond distribution over a heating path (that would give alower apparent equilibrium temperature estimated based on Δ47 mea-surement than the actual peak temperature) resulting for instancefrom tectonics (arc timescale regional heating) or from plutonic/dikeintrusion heating. However, our measured Δ47 values of Kamura

Page 8: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

KAMrm101KAMrm106KAMrm110

1 10 100 10000

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

B (mT)

f SIR

M

KAMrm13KAMrm16KAMrm 9

KAMrm13 (100 mT)KAMrm16 (100 mT)KAMrm 9 (100 mT)

KAMrm101 (100 mT)KAMrm106 (100 mT)KAMrm110 (100 mT)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

f IR

M10

0

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

f IR

M10

0

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

0 20 40 60 80 100 1200

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

B (mT)

frac

. mag

.

KAMrm13 ARM / IRMKAMrm16 ARM / IRMKAMrm 9 ARM / IRM

0 20 40 60 80 100 1200

0.2

0.4

0.6

0.8

1

B (mT)

fra

c. m

ag

.

KAMrm101 ARM / IRMKAMrm106 ARM / IRMKAMrm110 ARM / IRM

10−10

10−9

10−8

10−710

−12

10−11

10−10

10−9

10−8

IRM (Am2)

m A

m2

1:1000

1:1000

1:100

1:100

1:101:10

1:1KAM−13KAM−16KAM−09

r

−4

KAMrm101KAMrm106

KAMrm110

1 10 100 1000B (mT)

f SIR

M

B (mT)0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

B (mT)

10−10

10−9

10−8

10−7

IRM (Am2)

10−12

10−11

10−10

10−9

10−8

m A

m2

10−13

ARMsARMs

NRMs

NRMs

Fig. 4. Rock magnetic results from the Iwato and Mitai Formations, as described in the text. All results are consistent with fine-grained, presumably biogenic magnetite holding thecharacteristic remanent magnetization.

82 J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

Page 9: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

83J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

limestones are significantly higher than the minimum Δ47 value of0.350‰ observed in slowly cooled calcite marbles [corresponding tothe apparent equilibrium temperature or “blocking temperature” of190 °C for the Carrara marble standard, shown for comparison inTable 1]. This suggests that Kamura limestones probably never experi-enced temperatures higher than 190 °C for a long period of time (i.e.,longer than about 10–20 Myr based on parameters defined in Passeyand Henkes (2012)), because otherwise they would show Δ47 valuescomparable to Cararra marble. As noted in part 2.3 above, petrographicstudies of the large Alatoconchidae bivalves show that their externallayer retains an arguably primary prismatic calcite, with the originalbiological alignment of the c-axis perpendicular to the shell surface.Similar petrographic studies of the well-preserved micrites from theKamura limestone argue that a reasonable fraction of the carbonatehas not undergone obvious recrystallization. Higher temperaturesought to have led to petrographic textures and recrystallization typicalof marbles, which these rocks do not show. Finally, although some cau-tion iswarranted as knowledge on clumped isotopes in high temperaturesystems is still relatively recent, the most straightforward interpretationof our Δ47 results that are consistent to the known geological setting, isthat the actual peak heating temperature was close to 130 °C.

5.4. Paleomagnetic results

Specimen-level progressive demagnetization data for six samplesfrom three sections through the Kamura limestone are shown inFigs. S7A–F. The first thing to notice is the unusually weak nature of theNRM, with specimen magnetic moments starting in the range of ~10 to100 pAm2 (~10−8 to 10−7 emu, or about 2 to 20 μA/m for the 5 cm3

size of our typical specimen; remember that 1 G = 1 emu/cm3 =10+3 A/m). This is below holder noise on most of the superconductingmagnetometer systems in use, and a major reason that previouspaleomagnetic studies of rocks with similar lithologies have failed(Ripperdan, 1990; Yokoyama et al., 2007). Putting this in perspective,they are 2 to 3 orders ofmagnitudeweaker than the classic pelagic lime-stones of Italy, which have NRM magnetizations of ~1 mA/m (Alvarezand Lowrie, 1984). Nevertheless, the data are straightforward to inter-pret. Low-temperature cycling in liquid nitrogen (steps LT1 and LT2)removes up to ~10% of the NRM intensity, suggesting the presenceof occasional grains ofMDor SPDmagnetite that are unlockedwhile cy-cling through the Verway transition (Dunlop andOzdemir, 1997); theseparticles could be a remnant of background cosmic dust or volcanic ashin the environment that might be expected to fall into the sediment of acarbonate platform atoll. Directions for the component removed duringthis low-temperature cycling do not have an obvious grouping, andwere presumably gained randomly during sample transport or speci-men preparation. Following this, about half of the specimens display alow-coercivity, low blocking temperature overprint with directionssimilar to the Recent magnetic field direction, which is removed byweak AF demagnetization (up to 7 mT), and/or relatively low thermaltreatment (often to only 75 °C). Presumably, this is a combination ofviscous remanence, coupled with the presence of trace amounts of goe-thite formed during surface weathering (which generally carries a Re-cent or present field direction (PLF) which disappears as the mineralalters at low-temperature treatment). Most specimens at this pointeither retain or progressively move toward more characteristic rema-nence (ChRM) directions in two shallow inclination, antipodal groups,in the NNW and SSE directions. The increased scatter in the measure-ments is a result of theweak intensity of theNRMcompared to the com-bined background noise of the magnetometer and sample holders.

Oneproblem that arises commonly in paleomagnetic studies concernsthe recognition and separation of such soft (PLF) components from earli-er, characteristic (ChRM) directions. It helps if post-depositional tiltingof the units (and/or polar wander) creates a clear angular separation be-tween the components, although at best the twodirections canbeorthog-onal. The limestone block in the accretionary complex of the Kamura

region is tilted so that the bedding is now nearly vertical, but the axis ofthe mean PLF component removed by the initial low demagnetizationtreatment only has an angular separation of about 16° from the nearestaxis of the ChRM direction (that to the NNW; see below). This case istherefore similar to that of younger, flat-lying rocks where the distribu-tion of the two components also overlaps. We have therefore followedthe Principal Component procedure described by Tobin et al. (2012), inwhich we first examine data from samples that have the ChRM directionfurthest from the PLF direction, to gain an understanding of their relativestability spectra. We assume that the overprint is not influenced by thedirection of the underlying ChRM direction, and therefore use that as aguide for separating them during the Principal Component Analysis.As noted by Tobin et al., the scatter of the secular variation is oftenlarge enough so that clear ‘kinks’ appear in the vector demagnetizationdiagrams that allow separation of the components. This is discussed inmore detail in Fig. S7.

Results of the Principal Component Analysis (PCA) on the demagne-tization data from theKamura limestone are shown in Fig. 5 and Table 2.The soft component clusters around the expected direction for the Re-cent geomagnetic field, as indicated in Fig. 5A, and is identified as PLFin Table 2. In contrast, those specimens that contain a component ofhigher stability, the ‘Characteristic Remanence’ or ChRM, fall into twodistinct clusters. After correction for the local tilt of bedding these lieto theNNWwith generally shallow upward directions, and the oppositegroup is to the SSE with shallow down inclination (Fig. 5B). We empha-size that this ChRM direction is relatively weak compared to normal pa-leomagnetic studies. Using the matrix deconvolution method of Jones(2002), we have been able to estimate the total magnetic moment ofthe ChRMvector in some of our samples from the KAM locality (Section 8in Fig. 2), and plot them in Fig. 5C as a function of the geometric groupingparameter for linearity (MAD value, e.g., (Kirschvink, 1980)). Stabledirections are recoverable down to intensities of less than 10 pAm2

(10−8 emu). In a similar fashion, the ChRM component data can be ana-lyzed to yield an idea of the range of blocking temperatures that hold it bymaking a survivorship curve showing the last temperature step that pro-vides a useable point for the PCA analysis. As shown in Fig. 5D, these ther-mal termination points are distributed between ~200° and 400 °C, withabout 1/3 of the loss happening between 375° and 400 °C. Althoughwell below the Néel temperatures of the two major magnetic mineralspresent (magnetite and hematite, see Section 5.1 above), blocking tem-peratures are particle-size dependent and often fall in this range in natu-ral samples (such as bacterial magnetofossils.) This stability range is alsocompatible with the preservation of a primary remanence for fine-grained magnetite on geological time scales (Pullaiah et al., 1975).

Table 2 shows summary statistics of the PLF and ChRM components,the latter being given in coordinates of both in-situ and after correctionfor the local tilt of bedding, following the combined line and plane anal-ysis of McFadden and McElhinny (1988). Directions are also presentedseparating out samples according to polarity, as well as a grouping inwhich only data from specimens that yielded the least-squares linesare analyzed. Table S1 shows results of directional statistics tests.Watson's (1956) test of common Fisher's precision parameter (κ) indi-cates that the distributions are similar in both the NOR and REV cases,with or without the inclusion of demagnetization arcs, and that the κvalue for the PLF component is also similar to those for the ChRM. In ei-ther grouping (all data, or only demagnetization lines), the ChRM datashow a positive “Category B” reversals test, using the classification ofMcFadden & McEllhinny (McFadden and McElhinny, 1990). In particu-lar, the angular separation of the two antipolar groups using all of thedata is only 1.0°, whereas the critical angle for 95% confidence is 6.5°.Similarly, the axis of the PLF component, which lies 16.0° from theChRM axis, is statistically distinct from this at an extraordinarily highconfidence level. These data imply that the Kamura Limestone Atollformed at a paleolatitude of ~12°, and the pole position is far from anyother reported direction from the South China Block (e.g., (Wanget al., 1993; Yang and Besse, 2001)).

Page 10: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

90˚

180˚

270˚

N

E

S

W

A) Recent Overprint

C) ChRM Intensity

0

2

4

6

8

10

12

14

16

1 10 100 1000

MA

D v

alue

Intensity (pAm )2

B) ChRM Directions

50 100 150 200 250 300 350 400

Temperature (˚C)

120

100

80

60

40

20

0

ChR

Ms

Rem

aini

ng

D) ChRM Decay

Fig. 5. Results of the Principal Component Analysis. A. Low-coercivity and low-T components interpreted to be of Recent origin (PLF component, in-situ.). B. Two-polarity characteristicdirection, NOR and REV, tilt-corrected coordinates. Arc-constraints for samples that do not reach the stable endpoints are calculated with the method of McFadden and McElhinny(1988), and are shownwith the open symbols on the arc traces. Pink symbols indicate upper hemisphere directions. C. A summary of themaximum angular deviation from the principalcomponent analysis plotted against intensity of the fit principal component directions using thematrix deconvolution (J/Jo routine) of Jones (2002), described further in Kirschvink et al.(2008). D. Survivorship curve, showing the maximum temperature recorded for the ChRM component in each specimen as a function of temperature. For interpretation of the referencesto color in this figure legend, the reader is referred to the web version of this article.

84 J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

5.5. Magnetostratigraphy

Fig. 6 shows a summary comparison of themagnetic polarity resultsfor the Kamura limestones, as documented further in Supplemental

Table 2Paleomagnetic results from Upper Permian Limestones from Kyushu reported in this study.L = lines; P = planes; N = number of samples; Dec. = declination; Inc. = inclination; κ =pole latitude; VGP lon. = VGP longitude. *Values in these rows have been corrected for the tilThe field area in Kyushu, Japan is located approximately at 32.8° N, 131.3° E.

Components: N Dec.

PLF 105 354.1ChRMs

Normal lines 76 319.4* 327.0Reversed lines 43 152.8* 144.0Normal L + arcs 84 321.4* 327.9Reversed L + arcs 65.5 158.3* 147.0All (N + R) lines 119 145.1* 325.9

Best estimate of ChRM:(N + R) lines & arcs 149.5 149.6* 327.5

Pole position:Pole latitude: 35.3° A95:Pole longitude: 351.4° dp, dm:

Paleolatitude:

Figs. S7–9, and in the archived paleomagnetic data summary (MagICdata acquisition # XXXXX). The stratigraphically oldest interval, ex-posed in Section 2 (Saraito) starts in the N. craticulifera Zone (lowerWordian), and has a prominent switch in magnetic polarity between

Fisher's precision parameter; α9 = 95% confidence cone; VGP lat. = virtual geomagnetict of bedding.

Inc. κ R α95

55.5 15.20 98.2 3.7

69.1 11.12 69.3 5.1−24.4 12.86 70.2 4.7−62.5 15.12 40.2 5.822.7 14.83 40.2 5.969.5 11.48 76.8 4.8−23.6 13.27 77.7 4.4−61.7 14.89 61.2 4.723.1 14.81 61.1 4.7−66.8 12.03 109.2 3.9−23.8 13.56 110.3 3.7

−66.6 12.33 137.5 3.4−23.4 13.99 138.9 3.2

2.5°1.8°,3.5°−12.2 ± 1.8°

Page 11: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

100

90

80

70

60

50

40

30

20

10

0

120 m

110

N. craticulifera

N. margaritae

Yabeina

Lepidolina

barren

Codonofusiella-Reichelina

Lopingian

Lopi

ngia

n

Lopi

ngia

n

Gua

dalu

pian

Gua

dalu

pian

~26

5 m

a~

260

ma

Cis

ur-

alia

n

Cap

itani

an

Capitanian

Capitanian

Cap

itani

anW

uchi

apin

gian

Wuc

hiap

ingi

anW

ordi

an

Wordian

Wordian

Wor

dian

Roa

dian

Roadian

Sec.10

Section 8, top

Section 8 bottom

Sec. 6Sec. 3

Section 2 Bottom

Section 2 top

R

R

N

N

N

N

R

R

Sect. 9N

R

Permian Limestones, Kamura, Japan (Steiner, 2006)

main extinction

Kam

ura

even

t

+2 +4 +6Ccarb

13 ‰

+2 +4 +6

Ccarb13 ‰

Fig. 6.Magnetostratigraphy of the Permian Limestones of Kamura, Kyushu, Japan, compared with biostratigraphic and carbon isotope constraints. Data for the major sections are in thesupplemental information, Figs. S8–10. Note that the summary ignores occasional one-point samples with opposite polarity. Data for the Kamura carbon isotope anomaly are from(Isozaki, 1997a,b).

85J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

~10 and 12m in the section. This is approximately the position inferredfor the top of the Permian Long Reversed Chron, known as the “Illawara”reversal at the top of the Kiaman Reversed Interval (e.g., (Steiner,2006)). This implies that the North-Westerly and shallow-up polarityof the ChRM directions are of Normal polarity, and the South-Easterlyand shallow down directions are Reversed.

Using this polarity interpretation, the Wordian (N. craticulifera andN. margaritae Zones) sampled at Section 2 (Fig. S9) is characterizedby a clear reversed interval in the bottom 11 m of section, followedby an interval where 7 of the next 10 samples appear to be of normalpolarity (unfortunately separated by a 15 m thick covered interval).The Capitanian Yabeina and Lepidolina zones are characterized by solidnormal interval that is easily correlated with that reported from theCapitanian from the rest of the world (Gradstein et al., 2012; Steiner,2006). On the other hand, the barren interval of the topmost Capitanianis dominated by a reverse interval. This interval is possibly correlatedwith the reverse interval recognized in the upper Capitanian J. grantiZone in South China (Shen et al., 2010).

This interpretation is consistent with the remaining portions of thestratigraphy, in which the top of Section 2 lies in the N. margaritaeZone (upper Wordian), and should be N/R. Section 10 in the upperYabeina Zone (lower Capitanian) has a R/N transition, with the N extend-ing into the lower Lepidolina Zone. This switches to dominantly Rthrough the basal portion of the ‘barren’ interval in section 8 (uppermost

Capitanian) and extends well into Lopingian time (Codonofusiella–Reichelina Zone of the Wuchiapingian). Smaller stratigraphic intervals ofSections 2, 3, 6, and 9 are consistent with this interpretation, as shownin Fig. 6.

This polarity interpretation constrains the paleolatitude of thePermian Iwato and Mitai Limestone at Kamura to be at −12.2 ± 1.8°South, as indicated in Table 2.

6. Discussion

6.1. Stability and origin of the NRM

Although they possess an unusually weak magnetization, severallines of evidence support the hypothesis that samples of the late Permianlimestone from the Kamura area possess a primary remanent magneti-zation. Unrecrystallized calcite in the molluscan shell nacre and theclumped isotope results are most consistent with peak temperaturesbelow ~130 °C. In addition, the positive reversals test on the ChRMcomponent, its thermal stability spectrum, and the generally goodagreement of the reversal stratigraphy with other late Permian resultsargue for a primary, or early diagenetic origin of this component,most likely held in fine-grained magnetite. Coupled with the excellentbiostratrigraphic and chemostratigraphic constraints on the sequence,the paleomagnetic pole position calculated from this sequence scores

Page 12: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

86 J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

a perfect 7/7 on the Vandervoo (Vandervoo, 1990) scale of paleomag-netic reliability.

Other than a light scattering of cosmogenic or volcanic dust, most ofthe magnetization appears to be held in moderately interacting, veryfine-grained, single-domain or pseudo-single-domain magnetite. Themost likely origin of this material is production by the magnetotacticbacteria, as they are known to inhabit similar carbonate-rich environ-ments (Chang and Kirschvink, 1989; Chang et al., 1987; Kirschvinkand Chang, 1984; Kirschvink and Lowenstam, 1979; Kopp andKirschvink, 2008; Maloof et al., 2007) and have similar rock magneticproperties, including the inter-grain interaction effects (Kobayashiet al., 2006). They have also been implicated in the magnetization ofplatform carbonates from younger ocean island atolls (e.g., (Aissaouiand Kirschvink, 1991; Aissaoui et al., 1990)), as well as in a variety ofpelagic marine carbonates (Roberts et al., 2013).

An interesting twist on this result is the rockmagnetic evidence fromthe composite IRMexperiments (Figs. S1–5) for the presence of a signif-icant fraction of detrital hematite in sediments of the Iwato formationmentioned in Section 5.1 above. This unit preserves significantly lessorganicmatter than the underlyingMitai formation, and the associationof the hematite with the positive excursion of δ13Ccarb argues for amechanistic link. We suggest that the spike in δ13Ccarb actually reflectsan increase in the fraction of volcanic carbon (emitted as CO2) beingburied as organic carbon, with an implied 1–1 rise in atmospheric oxy-gen concentration. For shallow-water carbonates, an increased oxygenlevel would account for themore oxidized character of these sediments

6.2. Middle–Upper Permian magnetostratigraphy and bio- andchemostratigraphic correlation

Our data reveal multiple geomagnetic polarity changes withinthe Guadalupian Iwato Formation primarily from low-latitude mid-Panthalassa. As summarized in Fig. 6, the lower half (Wordian part) ofthe Iwato Formation is dominated by reversely magnetized intervals,whereas the upper half (Capitanian part) by normal ones. In contrast,the lower part of the overlyingMitai Formation above the G–LB recordeda stable reversed interval. This overall aspect is in general agreementwithprevious compilations of Permian magnetostratigraphy (e.g., (Gradsteinet al., 2012; Steiner, 2006)).

It is particularly noteworthy that a short normal interval appearsaround the G–LB (Fig. 6). This signal is possibly correlated with themagnetostratigraphic record from South China that is characterized bya short-term normal interval for the Clarkina postbitteri honshuiensisZone that defines the topmost Capitanian at the GSSP of G–LB in SouthChina (Shen et al., 2010). This correlation is supported by the stablecarbon isotope stratigraphy, because a sharp negative shift for ca. 3‰(from +5 to +2‰) was reported from the upper part of the barreninterval (Isozaki et al., 2007b). This shift is chemostratigraphically corre-latedwith the same signature in the C. postbitteri honshuiensis Zone andC. postbitteri postbitteri Zone across the G–LB at the GSSP in Penglaitan(Chen et al., 2011; Wang et al., 2004).

Similarly, the consistently reversely magnetized interval in thelowermost Codonofusiella–Reichelina Zone of the Mitai Formation alsofits into the reverse interval reported from the topmost C. postbitterihonshuiensis Zone, C. postbitteri postbitteri Zone, and the lowermostC. dukoensis Zone across the G–LB at the GSSP in South China (Shenet al., 2010).

These data add validity not only to the biostratigraphic correlationbased on fusuline zones in the paleo-atoll sections in Japan (Ota andIsozaki, 2006) and conodont zones in shelf sequences in South China(Shen and Mei, 2010), but also to the chemostratigraphical correlationbased on stable carbon isotope ratio of carbonates (Isozaki, 1997a,b).The Yabeina and Lepidolina zones of the Iwato Formation are correlatedwith the Capitanian conodont zones; i.e. the Jinogondolella postserrataZone to J. granti Zone, whereas the barren interval of the uppermostIwato Formation likely corresponds to the C. postbitteri honshuiensis

Zone that marks the uppermost Capitanian conodont zone, in SouthChina.

6.3. Identifying the Illarawa Reversal

From the viewpoint of magnetostratigraphy, it is particularly note-worthy that the lower part of the Iwato Formation is characterized bythe presence of samples with clear normal polarity in theN. craticulifera Zone of early Wordian age, which continues up into theoverlying N. margaritae Zone of late Wordian age. These data indicatethat a normal interval surely appeared in the Wordian, probably forthe first time since the Late Carboniferous. A long-lasting “reverse inter-val” that continued for ca. 50 million years from the Late Carboniferousto the Middle Permian has been widely known as the Kiaman ReverseSuperchron (Irving and Parry, 1963). The ‘Illawarra Reversal’ is an infor-mal termused tomark the top of the KiamanReverse Superchron, or thebase of the Permian–Triassic Mixed Superchron (Gradstein et al., 2012;Irving and Parry, 1963; Isozaki, 2009; Steiner, 2006). Nonetheless, itsage has not been precisely determined because it was originally detect-ed in the non-marine coal measures in eastern Australia that lack diag-nostic marine index fossils for global correlation. Therefore, previoussummaries for the Permian magnetostratigraphy were not necessarilyconsistent fromeach other, particularly for theWordian interval. For ex-ample, Ogg et al. (2008) placed the Illawarra Reversal at the base of theCapitanian, whereas Steiner (2006) and Henderson et al. (2012) put itin the middle of theWordian. Unfortunately, there are numerous prob-lems in compiling the late Permian information, as much of the datadiscussed by Steiner (2006) are unpublished, and many of the sectionsin Europe and Africa are in continental facies, where the correlations tothe marine realm are uncertain(Szurlies, 2013; Ward et al., 2005). Ourdata clearly show Normal chrons both in mid-Wordian time and mid tolate Capitanian time, with good biostratigraphic and chemostratigraphicconstraints (Ota and Isozaki, 2006).

The present data confirm that the Capitanian part (the Yabeina Zoneand Lepidolina Zone) possesses normal intervals. As to possible normalsub-chrons in the Wordian, however, two interpretations are possible:1) these signals can be regarded as small-scale geomagnetic noisewith-in the Kiaman Reverse Superchron, and the Illawarra Reversal in mid-superocean is stratigraphically placed around the G–LB as Ogg et al.(2008) proposed, or 2) one of them represents the first appearance ofa solid normal interval that pins down the Illawarra Reversal in mid-Wordian time. Judging from the frequent appearance of normal signals,we prefer the latter interpretation here, in accordance with the compi-lations of Steiner (2006) and Henderson et al. (2012).

6.4. Migration history of seamount and paleobiogeographical implications

The present paleomagnetic data demonstrate a long travel distanceof the Permian mid-superoceanic seamount capped by a paleo-atollcomplex. Around the G–L boundary, ~260 Ma, the seamount was accu-mulating shallowmarine atoll carbonates at 12.2 ± 1.8° S in themiddleof the Panthalassic ocean. The seamount later migrated toward north-west to eventually accrete to the Japan (= South China) marginat mid-latitude in the northern hemisphere during the mid-Jurassic(ca. 165–160 Ma), as indicated schematically in Fig. 7. The paleomag-netic data (and polarity interpretation) imply that it was subjectedto a vertical-axis rotation (w.r.t. North) of only about 30° CCW, overits entire history. The total travel distance of the seamount within west-ern Panthalassa likely reached ~3000 km, assuming a typical plateconvergence rate of 3 cm/year, and 100 myr to dock. During the north-wardmigration of the seamount, it likely crossed a paleo-biogeographicprovincial boundary, i.e. a border between the two contemporaryfusuline territories (e.g., Kasuya et al. (2012)). In fact, Kasuya et al.(2012) were the first to document the practical location of a mid-oceanic biogeographic province boundary within the pre-Jurassic lostoceans.We emphasize here that similar paleomagnetic studies onweakly

Page 13: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

Mid-Jurassic 160 Ma

S.China

60˚N

60˚S

30˚S

30˚N

equator

accretion

Pan

geaPanthalassa

Neo-Tethys Kamurapaleo-seamount

Shelf (-200–0 m)Deep Ocean (< -200 m)

Capitanian 260 Ma

Land (> 0 m)

equator

Pan

gea

Panthalassa

Panthalassa

Paleo-TethysS.China

Indochina

30˚N

60˚N

30˚S

60˚S

N.China

Sibumasu

Lepidolina territory

Yabeina territory

Yabeina territory

LepidolinaYabeina

Permianatoll

Fig. 7. Paleogeographic maps showing location of the Kamura seamount in late Permian and Jurassic time. Adapted from Kasuya et al. (2012).

87J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

magnetized paleo-atoll carbonates on migrating seamounts should havemany further applications in paleobiogeography, in particular for the lostwide oceanic domains in the past.

Acknowledgments

We thank Tomomi Kani of Kumamoto University, Masafumi Saitoh,Tomohiko Sato, and Daisuke Kofukuda of the University of Tokyo(Komaba) for their help in field work, and John Eiler of the CaliforniaInstitute of Technology for use of mass spectrometer for clumped iso-tope measurements.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.palaeo.2014.10.037.

References

Abrajevitch, A., Kodama, K., 2009. Biochemical vs. detrital mechanism of remanence ac-quisition inmarine carbonates: a lesson from the K–T boundary interval. Earth Planet.Sci. Lett. 286, 269–277.

Abrajevitch, A., Hori, R.S., Kodama, K., 2011. Magnetization carriers and remagnetizationof bedded chert. Earth Planet. Sci. Lett. 305, 135–142.

Aissaoui, D.M., Kirschvink, J.L., 1991. Atoll magnetostratigraphy — calibration of theEustatic records. Terra Nova 3, 35–40.

Aissaoui, D.M., McNeill, D.F., Kirschvink, J.L., 1990. Magnetostratigraphic dating ofshallow-water carbonates from Mururoa atoll, French Polynesia: implications forglobal eustasy. Earth Planet. Sci. Lett. 97, 102–112.

Alvarez, W., Lowrie, W., 1984. Magnetic stratigraphy applied to synsedimentary slumps,turbidites, and basin analysis; the Scaglia Limestone at Furlo (Italy). Geol. Soc. Am.Bull. 95, 324–336.

Alvarez, W., Kent, D.V., Silva, I.P., Schweickert, R.A., Larson, R.A., 1980. Franciscan complexlimestone deposited at 17° south paleolatitude. Geol. Soc. Am. Bull. 91, 476–484.

Ando, A., Kodama, K., Kojima, S., 2001. Low-latitude and Southern Hemisphere originof Anisian (Triassic) bedded chert in the Inuyama area, Mino terrane, central Japan.J. Geophys. Res. Solid Earth 106, 1973–1986.

Bonifacie, M., Ferry, J.M., Horita, J., Vasconcelos, C., Passey, B.H., Eiler, J.M., 2011.Calibration and applications of the dolomite clumped isotope thermometer to hightemperatures. Mineral. Mag. 75, 551.

Page 14: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

88 J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

Butler, R.F., 1992. Paleomagnetism: Magnetic Domains to Geologic Terranes. BlackwellScientific Publications, Boston.

Chang, S.-B.R., Kirschvink, J.L., 1989. Magnetofossils, the magnetization of sediments,and the evolution of magnetite biomineralization. Ann. Rev. Earth Planet. Sci. 17,169–195.

Chang, S.-B.R., Kirschvink, J.L., Stolz, J.F., 1987. Biogenic magnetite as a primary remanencecarrier in limestone. Phys. Earth Planet. Inter. 46, 289–303.

Chen, B., Joachimski, M.M., Sun, Y.D., Shen, S.Z., Lai, X.L., 2011. Carbon and conodontapatite oxygen isotope records of Guadalupian–Lopingian boundary sections: climaticor sea-level signal? Palaeogeogr. Palaeoclimatol. Palaeoecol. 311, 145–153.

Cisowski, S., 1981. Interacting vs. non-interacting single-domain behavior in natural andsynthetic samples. Phys. Earth Planet. Inter. 26, 56–62.

Cottrell, R.D., Tarduno, J.A., Roberts, J., 2008. The Kiaman reversed polarity superchron atKiama: toward a field strength estimate based on single silicate crystals. Phys. EarthPlanet. Inter. 169, 49–58.

Courtillot, V., Olson, P., 2007. Mantle plumes link magnetic superchrons to phanerozoicmass depletion events. Earth Planet. Sci. Lett. 260, 495–504.

Dekkers, M.J., 1989. Magnetic-properties of natural pyrrhotite 2. High-temperatureand low-temperature behavior of Jrs and TRM as a function of gain-size. Phys.Earth Planet. Inter. 57, 266–283.

Dunlop, D.J., Ozdemir, O., 1997. RockMagnetism: Fundamentals and Frontiers. CambridgeUniversity Press, New York.

Eiler, J.M., 2011. Paleoclimate reconstruction using carbonate clumped isotope thermom-etry. Quat. Sci. Rev. 30, 3575–3588.

Epstein, A.G., Epstein, J.L., Harris, L.D., 1977. Conodont color alteration — an index toorganic metamorphism. Geol. Soc. Am. Spec. Pap. 153, 108.

Fuller, M., Goree,W.S., Goodman, W.L., 1985. An introduction to the use of SQUIDmagne-tometers in biomagnetism. In: Kirschvink, J.L., Jones, D.S., MacFadden, B.J. (Eds.), Mag-netite Biomineralization and Magnetoreception in Organisms: a New Biomagnetism.Plenum Press, New York, pp. 103–151.

Fuller, M., Cisowski, S., Hart, M., Haston, R., Schmidke, E., Jarrard, R., 1988. NRM:IRM(s) demagnetization plots; an aid to the interpretation of natural remanent mag-netization. Geophys. Res. Lett. 15, 518–521.

Ghosh, P., Adkins, J., Affek, H., Balta, B., Guo, W.F., Schauble, E.A., Schrag, D., Eiler, J.M.,2006. 13C–18O bonds in carbonate minerals: a new kind of paleothermometer.Geochim. Cosmochim. Acta 70, 1439–1456.

Gialanella, P.R., Heller, F., Haag, M., Nurgaliev, D., Borisov, A., Burov, B., Jasonov, P.,Khasanov, D., Ibragimov, S., Zharkov, I., 1997. Late Permian magnetostratigraphy onthe eastern Russian platform. Geol. En Mijnbouw 76, 145–154.

Gradstein, F., Ogg, J.G., Schmitz, M., Ogg, G., 2012. The Geological Time Scale 2012.Elsevier, Boston.

Guo, W., Mosenfelder, J.L., Goddard, W.A., Eiler, J.M., 2009. Isotopic fractionationsassociated with phosphoric acid digestion of carbonate minerals: insights from firstprinciples theoretical modeling and clumped isotope measurements. Geochim.Cosmochim. Acta 73, 7203–7225.

Henderson, C.M., Davydov, V.I., Wardlaw, B.R., 2012. The Permian period. In: Gradstein,F.M., Ogg, J.G., Schmitz, M.D., Ogg, G.M. (Eds.), The Geologic Time Scale 2012. Elsevier,Amsterdam, pp. 653–679.

Irving, E., 1964. Paleomagnetism and its Application to Geologic and GeophysicalProblems. Wiley, New York, New York.

Irving, E., Parry, L.G., 1963. The magnetism of some Permian rocks from New-South-Wales. Geophys. J. R. Astron. Soc. 7, 395–411.

Isozaki, Y., 1997a. Contrasting two types of orogen in Permo-Triassic Japan: accretionaryversus collisional. Island Arc 6, 2–24.

Isozaki, Y., 1997b. Jurassic accretion tectonics of Japan. Island Arc 6, 25–51.Isozaki, Y., 2006. Guadalupian (Middle Permian) giant bivalve Alatoconchidae from a

mid-Panthalassan paleo-atoll complex in Kyushu, Japan: a unique community associ-ated with Tethyan fusulines and corals. Proc. Jpn. Acad. Ser. B Phys. Biol. Sci. 82,25–32.

Isozaki, Y., 2009. Illawarra Reversal: the fingerprint of a superplume that triggeredPangean breakup and the end-Guadalupian (Permian) mass extinction. GondwanaRes. 15, 421–432.

Isozaki, Y., 2014. Memories of Pre-Jurassic Lost Oceans: how to retrieve them from extantlands. Geosci. Can. 41, 283–311.

Isozaki, Y., Aljinovic, D., 2009. End-Guadalupian extinction of the Permian gigantic bivalveAlatoconchidae: end of gigantism in tropical seas by cooling. Palaeogeogr.Palaeoclimatol. Palaeoecol. 284, 11–21.

Isozaki, Y., Ota, A., 2001. Middle–Upper Permian (Maokouan-Wuchiapingian) boundaryin mid-oceanic paleo-atoll limestone of Kamura and Akasaka, Japan. Proc. Jpn. Acad.Ser. B Phys. Biol. Sci. 77, 104–109.

Isozaki, Y., Maruyama, S., Furuoka, F., 1990. Accreted oceanic materials in Japan.Tectonophysics 181, 179–205.

Isozaki, Y., Kawahata, H., Minoshima, K., 2007a. The Capitanian (Permian) KamuraCooling Event: the beginning of the Paleozoic–Mesozoic transition. Palaeoworld 16,16–30.

Isozaki, Y., Kawahata, H., Ota, A., 2007b. A unique carbon isotope record acrossthe Guadalupian–Lopingian (Middle–Upper Permian) boundary in mid-oceanicpaleo-atoll carbonates: the high-productivity “Kamura event” and its collapse inPanthalassa. Glob. Planet. Chang. 55, 21–38.

Isozaki, Y., Aljinovic, D., Kawahata, H., 2011. The Guadalupian (Permian) Kamura event inEuropean Tethys. Palaeogeogr. Palaeoclimatol. Palaeoecol. 308, 12–21.

Jin, Y.G., Zhang, J., Shang, Q.H., 1994. Two phases of the end-Permian mass extinction. In:Embry, A.F., Beauchamp, B., Glass, D.J. (Eds.), Pangea: Global Environments andResourcesMemoir. Canadian Society of Petroleum Geologists, pp. 813–822.

Johnson, H.P., Lowrie, W., Kent, D.V., 1975. Stability of ARM in fine and course grainedmagnetite and maghemite particles. Geophys. J. R. Astron. Soc. 41, 1–10.

Jones, C.H., 2002. User-driven integrated software lives: “Paleomag” paleomagneticsanalysis on the Macintosh. Comput. Geosci. 28, 1145–1151.

Kambe, N., 1963. On the boundary between the Permian and Triassic Systems in Japanwith the description of the Permo–Triassic formations at Takachiho-cho, MiyazakiPrefecture in Kyushu and the Skytic fossils contained. Geological Survey of JapanReport, pp. 1–68.

Kani, T., Fukui, M., Isozaki, Y., Nohda, S., 2008. The Paleozoicminimumof Sr-87/Sr-86 ratioin the Capitanian (Permian) mid-oceanic carbonates: a critical turning point in theLate Paleozoic. J. Asian Earth Sci. 32, 22–33.

Kani, T., Hisanabe, C., Isozaki, Y., 2013. The Capitanian (Permian) minimum of Sr-87/Sr-86ratio in the mid-Panthalassan paleo-atoll carbonates and its demise by the deglacia-tion and continental doming. Gondwana Res. 24, 212–221.

Kanmera, K., Nakazawa, K., 1973. Permian–Triassic relationships and faunal changesin the eastern Tethys. In: Logan, A., Hills, L.V. (Eds.), The Permian and TriassicSystems and their Mutual Boundary. Canadian Society of Petroleum GeologistsMemoir, pp. 100–119.

Kasuya, A., Isozaki, Y., Igo, I., 2012. Constraining paleo-latitude of a biogeographic bound-ary in mid-Panthalassa: Fusuline province shift on the Late Guadalupian (Permian)migrating seamount. Gondwana Res. 21, 611–623.

Kirschvink, J.L., 1980. The least-squares line and plane and the analysis of paleomagneticdata: examples from Siberia and Morocco. Geophys. J. R. Astron. Soc. 62, 699–718.

Kirschvink, J.L., 1981. How sensitive should a rock magnetometer be for use in paleomag-netism? In: Weinstock, H., Overton, W.C. (Eds.), SQUID Applications to Geophysics.The Society of Exploration Geophysicists, Tulsa, Oklahoma, pp. 111–114.

Kirschvink, J.L., 1992. Uniform magnetic fields and Double-wrapped coil systems:improved techniques for the design of biomagentic experiments. Bioelectromagnetics13, 401–411.

Kirschvink, J.L., Chang, S.-B.R., 1984. Ultra fine-grained magnetite in deep-sea sediments:possible bacterial magnetofossils. Geology 12, 559–562.

Kirschvink, J.L., Lowenstam, H.A., 1979. Mineralization and magnetization of chiton teeth:paleomagnetic, sedimentologic, and biologic implications of organic magnetite. EarthPlanet. Sci. Lett. 44, 193–204.

Kirschvink, J.L., Kopp, R.E., Raub, T.D., Baumgartner, C.T., Holt, J.W., 2008. Rapid, precise,and high-sensitivity acquisition of paleomagnetic and rock-magnetic data: develop-ment of a low-noise automatic sample changing system for superconducting rockmagnetometers. Geochem. Geophys. Geosyst. 9, 1–18.

Kobayashi, A., Kirschvink, J.L., Nash, C.Z., Kopp, R.E., Sauer, D.A., Bertani, L.E., Voorhout,W.F., Taguchi, T., 2006. Experimental observation of magnetosome chain collapse inmagnetotactic bacteria: sedimentological, paleomagnetic, and evolutionary implica-tions. Earth Planet. Sci. Lett. 245, 538–550.

Kodama, K., Takeda, T., 2002. Paleomagnetism ofmid-Cretaceous red beds inwest-centralKyushu Island, southwest Japan: paleoposition of Cretaceous sedimentary basinsalong the eastern margin of Asia. Earth Planet. Sci. Lett. 201, 233–246.

Koike, T., 1996. The first occurrence of Griesbachian conodonts in Japan. Transaction andProceedings of the Palaeontological Society of Japan, New Series 181, pp. 337–346.

Kopp, R.E., Kirschvink, J.L., 2008. The identification and biogeochemical interpretation offossil magnetotactic bacteria. Earth Sci. Rev. 86, 42–61.

Lowrie, W., 1990. Identification of ferromagnetic minerals in a rock by coercivity andunblocking temperature properties. Geophys. Res. Lett. 17, 159–162.

Maloof, A.C., Kopp, R.E., Grotzinger, J.P., Fike, D.A., Bosak, T., Vali, H., Poussart, P.M., Weiss,B.P., Kirschvink, J.L., 2007. Sedimentary Iron Cycling and the Origin and Preservationof Magnetization in Platform Carbonate Muds, Andros Island, the Bahamas. EarthPlanet. Sci. Lett. 259, 581–598.

McFadden, P.L., McElhinny, M.W., 1988. The combined analysis of remagnetization circlesand direct observations in Paleomagnetism. Earth Planet. Sci. Lett. 87, 161–172.

McFadden, P.L., McElhinny, M.W., 1990. Classification of the reversal test in paleomagne-tism. Geophys. J. Int. 103, 725–729.

Oda, H., Suzuki, H., 2000. Paleomagnetism of Triassic and Jurassic red bedded chert of theInuyama area, central Japan. J. Geophys. Res. Solid Earth 105, 25743–25767.

Ogg, J.G., Ogg, G., Gradstein, F.M., 2008. Concise Geologic Time Scale. CambridgeUniversityPress, Cambridge, U.K.

Opdyke, N.D., Roberts, J., Claoue-Long, J., Irving, E., Jones, P.J., 2000. Base of the Kiaman: itsdefinition and global stratigraphic significance. Geol. Soc. Am. Bull. 112, 1315–1341.

Ota, A., Isozaki, Y., 2006. Fusuline biotic turnover across the Guadalupian–Lopingian(Middle–Upper Permian) boundary inmid-oceanic carbonate buildups: biostratigraphyof accreted limestone in Japan. J. Asian Earth Sci. 26, 353–368.

Passey, B.H., Henkes, G.A., 2012. Carbonate clumped isotope bond reordering andgeospeedometry. Earth Planet. Sci. Lett. 351, 223–236.

Pullaiah, G., Irving, G., Buchan, K.L., Dunlop, D.J., 1975. Magnetization changes caused byburial and uplift. Earth Planet. Sci. Lett. 28, 133–143.

Ripperdan, R.L., 1990. Magnetostratigraphic investigations of the lower Paleozoic periodboundaries, and associated paleogeographic implications. Geological & PlanetarySciencesCalifornia Institute of Technology, Pasadena, California, p. 171.

Roberts, A.P., Florindo, F., Chang, L., Heslop, D., Jovane, L., Larrasoana, J.C., 2013. Magneticproperties of pelagic marine carbonates. Earth Sci. Rev. 127, 111–139.

Sano, H., Nakashima, K., 1997. Lowermost Triassic (Griesbachian) microbial bindstone-cementstone facies, southwest Japan. Facies 36, 1–24.

Seguin, M.K., Petryk, A.A., 1986. Paleomagnetic study of the Late Ordovician–Early Silurianplatform sequence of Anticosti Island, Quebec. Can. J. Earth Sci. 23, 1880–1890.

Shen, S.Z., Mei, S.L., 2010. Lopingian (Late Permian) high-resolution conodont biostratig-raphy in Iran with comparison to South China zonation. Geol. J. 45, 135–161.

Shen, S.Z., Henderson, C.M., Bowring, S.A., Cao, C.Q.,Wang, Y.,W.,W., Zhang, H., Zhang, Y.C.,Mu, L., 2010. High-resolution Lopingian (Late Permian) timescale of South China. Geol.J. 45, 122–134.

Shibuya, H., Sasajima, S., 1986. Paleomagnetism of red cherts — a case-study in theInuyama area, central Japan. J. Geophys. Res. Solid Earth Planets 91, 14105–14116.

Page 15: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

89J.L. Kirschvink et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 418 (2015) 75–89

Stanley, S.M., Yang, X., 1994. A double mass extinction at the end of the Paleozoic era.Science 266, 1340–1344.

Steiner, M.B., 2006. The magnetic polarity time scale across the Permian–Triassic bound-ary. In: Lucas, S.G., Casinis, G., Schneider, J.W. (Eds.), Non-marine Permian biostratig-raphy and biochronology. Geological Society of London, pp. 15–38.

Szurlies, M., 2013. Late Permian (Zechstein) magnetostratigraphy inWestern and CentralEurope. Geol. Soc. Lond. Spec. Publ. 376, 73–85.

Tobin, T.S., Ward, P.D., Steig, E.J., Olivero, E.B., Hilburn, I.A., Mitchell, R.N., Diamond, M.R.,Raub, T.D., Kirschvink, J.L., 2012. Extinction patterns, delta(18) O trends, andmagnetostratigraphy from a southern high-latitude Cretaceous–Paleogene section:links with Deccan volcanism. Palaeogeogr. Palaeoclimatol. Palaeoecol. 350, 180–188.

Uno, K., Onoue, T., Hamada, K., Hamami, S., 2012. Palaeomagnetism of Middle Triassic redbedded cherts from southwest Japan: equatorial palaeolatitude of primary magneti-zation and widespread secondary magnetization. Geophys. J. Int. 189, 1383–1398.

Vandervoo, R., 1990. The reliability of paleomagnetic data. Tectonophysics 184, 1–9.Wang, Z.M., Vandervoo, R., Wang, Y.G., 1993. A new Mesozoic apparent polar wander

loop for South China — Paleomagnetism of Middle Triassic rocks from GuizhouProvince. Earth Planet. Sci. Lett. 115, 1–12.

Wang, W., Cao, C.Q., Wang, Y., 2004. The carbon isotope excursion on GSSP candidatesection of Lopingian–Guadalupian boundary. Earth Planet. Sci. Lett. 220, 57–67.

Ward, P.D., Botha, J., Buick, R., De Kock, M.O., Erwin, D.H., Garrison, G.H., Kirschvink, J.L.,Smith, R., 2005. Abrupt and gradual extinction among Late Permian land vertebratesin the Karoo Basin, South Africa. Science 307, 709–714.

Wardlaw, B.R., Davydov, V., Gradstein, F.M., 2004. The Permian period. In: Gradstein, F.M.,Ogg, J.M., Smith, A.G. (Eds.), A Geologic Time Scale 2004. Cambridge University Press,Cambridge, pp. 249–270.

Watson, G.S., 1956. Analysis of dispersion on a sphere. Geophys. J. R. Astron. Soc. 7,153–159.

Weiss, B.P., Baudenbacher, F.J., Wikswo, J.P., Kirschvink, J.L., 2001. Magnetic microscopypromises a leap in sensitivity and resolution. EOS Trans. Am. Geophys. Union 82,513–518.

Yang, Z.Y., Besse, J., 2001. New Mesozoic apparent polar wander path for south China:tectonic consequences. J. Geophys. Res. Solid Earth 106, 8493–8520.

Yokoyama, M., Isozaki, Y., Otofuji, Y.-I., 2007. Rock magnetic study of the Middle to LatePermian shallow-sea limestones in the central Kyusyu. Japan Geoscience UnionMeeting, Chiba City, Japan, p. E111-009.

Page 16: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

Supplementary Figures:

Fig. S1. Sampling locality for the rock-magnetic survey, along the road at

Section 8 (Fig. S10, below).

Fig. S2. IRM acquisition and backfield characteristics for representative

samples from the Mitai and Iwato formations.

Fig. S3. Stratigraphic variability in the IRM acquisition properties of (a) the

Mitai Fm, (b) the G/L boundary interval, and (c) the Iwato Fm.

Fig S4. Composite IRM thermal demagnetization showing the presence of

magnetite only, and magnetite + hematite, in samples of the Iwato and Mitai

formations.

Fig. S5. IRM acquisition and thermal demagnetization of composite-IRMs

experiments from the Iwato formation, and across the G/L boundary interval.

Fig. S6. SQUID magnetic microscope image of the NRM over the surface of

specimen KAM-109.2, from the upper portion of the Iwato formation. The image

was made in steps of 75 um, over a surface area of 26.5 x 28 mm, at a vertical offset

distance (from the superconducting probe) of about 250 um. A faint ghost image of

the edge of the 2.54 cm diameter sample is clearly visible, which is a result of the

abrupt boundary from the downwards-magnetized sample over free space. A few

minor anomalies suggest local moments on the order of a pAm2, accounting for at

most 10% of the NRM of the specimen. Vertical stripes in the data are the result of

instrument drift, reflecting the fact that the sample is extraordinarily weakly

magnetized, even approaching the limit of detection of the SQUID microscope

system.

Fig. S7. Example demagnetization data from the weakly magnetic Permian

carbonates from Kamura, Kyushu, Japan (Isozaki et al., 2007b). Data are shown

both as orthogonal projections (with in-situ, geographic coordinates, left-side

diagrams), as well as portions of equal-area projections (corrected for local tilt of

bedding, right-side diagrams). A subset of points are labeled according to the

demagnetization treatment as follows: the NRM is the first vector measured after

sample preparation and several weeks storage in the magnetically-shielded room;

thermal cycling to 77 K in liquid nitrogen is indicated by LTn, where n is the

number of times the sample was cycled; the peak level of 3-axis alternating-field

treatment is given by AFn.n, where n.n is the peak field in mT; and the peak

Page 17: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

temperature in ˚C is shown from thermal demagnetization in N2 gas. On the

orthogonal projections, the pair of light-grey arrows shows the approximate range of

the PLF component, and the green arrows that of the ChRM. Specimen A (Sara

23.1) shows the removal of a strong PLF component between the Af6.9 mT step and

the first heating to 75˚C, with a sharp angular separation with the reversely-

magnetized ChRM direction. In B (Sara 36.1), the PLF direction displays slightly

higher stability, with the angular break between it and the normal polarity ChRM

direction happening at about 200˚C; however, the angular break is clearly

discernable. Sample C (KAM 12.1) is an example where the PLF and the ChRM

directions have less of an angular separation, but can still be distinguished easily by

a slight kink in the demagnetization trajectory. Samples in D, E and F are more

ideal, where the magnitude of the PLF component relative to ChRM direction is

minimal.

Fig. S8. Magnetostratigraphy of the Iwato formation at Section 10

(Hijirigawa). Lithostratigraphic colums are after Isozaki et al. (Isozaki et al., 2011;

Isozaki et al., 2007a), and the stratigraphic position of samples that yielded ChRM

components are shown. Data points show the declination of the ChRM component

after correction for the tilt of the bedding, with the solid symbols indicating those

that gave linear components. Open symbols show the declination of the best-fit

iterated direction from specimens that followed great-circle arcs, using the iterative

approach of McFadden and McElhinny (1988). Polarity interpretations are as

indicated in the text, and the lithostratigraphic legend is given on Fig. S8.

Fig. S9. Magnetostratigraphy of the Iwato formation at Section 2 (Saraito).

Symbols are as described on Fig. S8.

Fig. S10. Magnetostratigraphy of the Iwato and Mitai formation at Section 8

(Shioinouso, the ‘KAM’ samples). Symbols are as described on Fig. S8.

Page 18: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

boundary

N

a)

-8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

sampling position (m)b)boundaryIwato Fm. Mitai Fm.

Fig.1

Page 19: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

Mitai Fm.

Iwato Fm.

-2000 200 600 1000 1400 1800 2200 26000.0E+05.0E-21.0E-11.5E-12.0E-12.5E-13.0E-13.5E-14.0E-14.5E-15.0E-15.5E-16.0E-16.5E-1

MT1100-3

Applied field (mT)

IRM

(A/m

)

-2000 200 600 1000 1400 1800 2200 26000.0E+0

1.0E-1

2.0E-1

3.0E-1

4.0E-1

5.0E-1

6.0E-1

7.0E-1

8.0E-1

9.0E-1

1.0E+0

1.1E+0MT0000-1

Applied field (mT)

IRM

(A/m

)

-2000 200 600 1000 1400 1800 2200 26000.0E+0

2.5E-2

5.0E-2

7.5E-2

1.0E-1

1.3E-1

1.5E-1

1.8E-1

2.0E-1

2.3E-1

2.5E-1

2.8E-1IW0025-1

Applied field (mT)

IRM

(A/m

)

-2000 200 600 1000 1400 1800 2200 26000.0E+01.0E-22.0E-23.0E-24.0E-25.0E-26.0E-27.0E-28.0E-29.0E-21.0E-11.1E-11.2E-11.3E-11.4E-11.5E-11.6E-1

IW0050-1

Applied field (mT)

IRM

(A/m

)

-2000 200 600 1000 1400 1800 2200 26000.0E+0

2.5E-2

5.0E-2

7.5E-2

1.0E-1

1.3E-1

1.5E-1

1.8E-1

2.0E-1

2.3E-1IW0600-1

Applied field (mT)

IRM

(A/m

)

-2000 200 600 1000 1400 1800 2200 26000.0E+0

5.0E-3

1.0E-2

1.5E-2

2.0E-2

2.5E-2

3.0E-2

3.5E-2

4.0E-2

4.5E-2

5.0E-2MT0400-3

Applied field (mT)IR

M (A

/m)

Fig.2

Page 20: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

-8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+05.0E-21.0E-11.5E-12.0E-12.5E-13.0E-13.5E-14.0E-14.5E-15.0E-15.5E-16.0E-16.5E-1

MT1100-3

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+01.0E-22.0E-23.0E-24.0E-25.0E-26.0E-27.0E-28.0E-29.0E-21.0E-11.1E-11.2E-1

MT0700-3

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+0

5.0E-3

1.0E-2

1.5E-2

2.0E-2

2.5E-2

3.0E-2

3.5E-2

4.0E-2

4.5E-2

5.0E-2MT0400-3

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+0

2.0E-2

4.0E-2

6.0E-2

8.0E-2

1.0E-1

1.2E-1

1.4E-1

1.6E-1

1.8E-1MT0800-1

Applied Field (mT)IR

M (A

/m)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+02.5E-25.0E-27.5E-21.0E-11.3E-11.5E-11.8E-12.0E-12.3E-12.5E-12.8E-13.0E-13.3E-13.5E-1

MT0500-2

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+0

1.0E-2

2.0E-2

3.0E-2

4.0E-2

5.0E-2

6.0E-2

7.0E-2

8.0E-2

9.0E-2MT0300-3

Applied Field (mT)

IRM

(A/m

)

Mitai Fm.Fig.3 a)

Page 21: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

-8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+0

2.0E-2

4.0E-2

6.0E-2

8.0E-2

1.0E-1

1.2E-1

1.4E-1

1.6E-1MT0100-2

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+01.0E-12.0E-13.0E-14.0E-15.0E-16.0E-17.0E-18.0E-19.0E-11.0E+01.1E+0

MT0000-1

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+0

2.0E-2

4.0E-2

6.0E-2

8.0E-2

1.0E-1

1.2E-1

1.4E-1

1.6E-1IW0050-1

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+01.0E-22.0E-23.0E-24.0E-25.0E-26.0E-27.0E-28.0E-29.0E-21.0E-11.1E-11.2E-1

MT0050-2

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+02.5E-25.0E-27.5E-21.0E-11.3E-11.5E-11.8E-12.0E-12.3E-12.5E-12.8E-1

IW0025-1

Applied Field (mT)

IRM

(A/m

)

around the boundary

Fig.3 b)

Page 22: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

-8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+01.0E-12.0E-13.0E-14.0E-15.0E-16.0E-17.0E-18.0E-19.0E-11.0E+01.1E+01.2E+0

IW0250-1

Applied Field (mT)IR

M (A

/m)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+0

2.5E-2

5.0E-2

7.5E-2

1.0E-1

1.3E-1

1.5E-1

1.8E-1

2.0E-1

2.3E-1IW0600-1

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+01.0E-22.0E-23.0E-24.0E-25.0E-26.0E-27.0E-28.0E-29.0E-21.0E-11.1E-1

IW0800-2

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+0

2.0E-1

4.0E-1

6.0E-1

8.0E-1

1.0E+0

1.2E+0

1.4E+0

1.6E+0

1.8E+0IW0150-2

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+0

1.0E-1

2.0E-1

3.0E-1

4.0E-1

5.0E-1

6.0E-1

7.0E-1

8.0E-1

9.0E-1

1.0E+0IW0500-1

Applied Field (mT)

IRM

(A/m

)

0 200 400 600 8001000 1400 1800 2200 2600

0.0E+0

2.5E-2

5.0E-2

7.5E-2

1.0E-1

1.3E-1

1.5E-1

1.8E-1

2.0E-1

2.3E-1IW0650-1

Applied Field (mT)

IRM

(A/m

)

Iwato Fm.Fig.3 c)

Page 23: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

0 100 200 300 400 500 600 7000.0E+0

2.0E-2

4.0E-2

6.0E-2

8.0E-2

1.0E-1

1.2E-1

1.4E-1

1.6E-1

1.8E-1

2.0E-1

IW0025-1

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 7000.0E+0

1.0E-2

2.0E-2

3.0E-2

4.0E-2

5.0E-2

6.0E-2

7.0E-2

8.0E-2

9.0E-2

1.0E-1

1.1E-1

IW0050-1

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 7000.0E+01.0E-22.0E-23.0E-24.0E-25.0E-26.0E-27.0E-28.0E-29.0E-21.0E-11.1E-11.2E-11.3E-1

MT0100-2

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 7000.0E+05.0E-31.0E-21.5E-22.0E-22.5E-23.0E-23.5E-24.0E-24.5E-25.0E-25.5E-26.0E-26.5E-27.0E-27.5E-2

MT0300-3

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 7000.0E+0

5.0E-2

1.0E-1

1.5E-1

2.0E-1

2.5E-1

3.0E-1

3.5E-1

4.0E-1

4.5E-1

5.0E-1

5.5E-1

IW0500-1

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 7000.0E+0

1.0E-2

2.0E-2

3.0E-2

4.0E-2

5.0E-2

6.0E-2

7.0E-2

8.0E-2

9.0E-2

1.0E-1

1.1E-1

IW0600-1

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

Magnetite only

Magnetite+Hematite

Fig.4 a)

Page 24: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

0 100 200 300 400 500 600 7000.0E+0

2.0E-1

4.0E-1

6.0E-1

8.0E-1

1.0E+0

1.2E+0

1.4E+0

1.6E+0

1.8E+0

IW0150-2

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 7000.0E+05.0E-21.0E-11.5E-12.0E-12.5E-13.0E-13.5E-14.0E-14.5E-15.0E-15.5E-16.0E-1

IW0250-1

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 7000.0E+0

1.0E-2

2.0E-2

3.0E-2

4.0E-2

5.0E-2

6.0E-2

7.0E-2

8.0E-2

9.0E-2

IW0800-2

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

Magnetite+Geothite

Magnetite only

Fig.4 b)

Page 25: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

-8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

Mitai Fm.

0 100 200 300 400 500 600 700

0.0E+0

1.0E-2

2.0E-2

3.0E-2

4.0E-2

5.0E-2

6.0E-2

7.0E-2

8.0E-2MT0300-3

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 700

0.0E+0

5.0E-3

1.0E-2

1.5E-2

2.0E-2

2.5E-2

3.0E-2

3.5E-2

4.0E-2

4.5E-2

5.0E-2MT0400-3

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 700

0.0E+0

1.0E-2

2.0E-2

3.0E-2

4.0E-2

5.0E-2

6.0E-2

7.0E-2

8.0E-2

9.0E-2

1.0E-1MT0700-3

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 700

0.0E+0

5.0E-2

1.0E-1

1.5E-1

2.0E-1

2.5E-1

3.0E-1

3.5E-1

4.0E-1

4.5E-1

5.0E-1MT1100-3

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

Fig.5 a)

Page 26: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

-8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

0 100 200 300 400 500 600 700

0.0E+01.0E-22.0E-23.0E-24.0E-25.0E-26.0E-27.0E-28.0E-29.0E-21.0E-11.1E-1

IW0050-1

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 700

0.0E+0

1.0E-1

2.0E-1

3.0E-1

4.0E-1

5.0E-1

6.0E-1

7.0E-1

8.0E-1

9.0E-1MT0000-1

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 700

0.0E+0

2.5E-2

5.0E-2

7.5E-2

1.0E-1

1.3E-1

1.5E-1

1.8E-1

2.0E-1IW0025-1

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 700

0.0E+01.0E-22.0E-23.0E-24.0E-25.0E-26.0E-27.0E-28.0E-29.0E-21.0E-11.1E-11.2E-11.3E-1

MT0100-2

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

around the boundaryFig.5 b)

Page 27: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

-8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

0 100 200 300 400 500 600 700

0.0E+0

1.0E-2

2.0E-2

3.0E-2

4.0E-2

5.0E-2

6.0E-2

7.0E-2

8.0E-2

9.0E-2IW0800-2

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 700

0.0E+01.0E-22.0E-23.0E-24.0E-25.0E-26.0E-27.0E-28.0E-29.0E-21.0E-11.1E-1

IW0600-1

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 700

0.0E+05.0E-21.0E-11.5E-12.0E-12.5E-13.0E-13.5E-14.0E-14.5E-15.0E-15.5E-16.0E-1

IW0250-1

0.12T0.4T2.7T

Temperature (oC)IR

M (A

/m)

0 100 200 300 400 500 600 700

0.0E+05.0E-21.0E-11.5E-12.0E-12.5E-13.0E-13.5E-14.0E-14.5E-15.0E-15.5E-1

IW0500-1

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

0 100 200 300 400 500 600 700

0.0E+0

2.0E-1

4.0E-1

6.0E-1

8.0E-1

1.0E+0

1.2E+0

1.4E+0

1.6E+0

1.8E+0IW0150-2

0.12T0.4T2.7T

Temperature (oC)

IRM

(A/m

)

Iwato Fm.Fig.5 c)

Page 28: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

Distance (mm)

Dis

tanc

e (m

m)

Verti

cal M

agne

tic F

ield

(nT)

0 5 10 15 20 250

5

10

15

20

25

−20

−15

−10

−5

0

5

10

15

20

25

30

Page 29: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

LT1

LT2

AF6.9

NRM

NRMNRM

LT1LT2

AF6.9

AF6.9

AF6.9

120˚ 105˚

150˚

160˚170˚

Up, W

Up, NDn, E

EW

Horizonal orlower Hem.

Verticalor upper Hem.

A. Sara 23.1

B. Sara 36.1

S

NRM

NRM

LT2

LT2

AF2.3

105˚

105˚

260˚

260˚

75˚

75˚

90˚

250˚

250˚

300˚280˚

150˚

75˚

120˚

120˚

180˚

200˚

230˚

Tilt-corrected coordinates

Tilt-corrected coordinates

In-Situcoordinates

In-Situcoordinates

180˚

90˚ Up,Dn

1 pAm / Division 2

1 pAm / Division 2

LT1

300˚250˚

260˚

Up,Dn

Kirschvink et al., Fig. S7 A,B

Page 30: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

NRM

NRM

NRM

NRM

LT1

LT1

LT1

120˚

360˚

360˚

255˚

295˚

360˚

Up, N

Up, N

Dn, S

E

EW

Horizonal orlower Hem.

Verticalor upper Hem.

C. KAM12.1

D. KAM34.2

10 pAm / Division

1 pAm / Division

NRM

LT2

AF2.3

AF6.9 105˚

105˚

180˚

195˚

210˚225˚

180˚

210˚

105˚

75˚

75˚

Tilt-corrected coordinates

In-Situ coordinates

In-Situcoordinates

Up,Dn

2

2

W

180˚

90˚

Tilt-corrected coordinates

Kirschvink et al., Fig. S7 C,D

Page 31: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

AF6.9

NRM

AF2.3

AF2.3

AF4.6AF6.9

Up, N

Up, NDn, S

Dn, S

E

E

W

W

Horizonal orlower Hem.

Verticalor upper Hem.

E. HJPM-5.1

F. HJPM-a.1

NRM

NRM

NRM

LT2

LT1

LT1

AF6.9 75˚

75˚

75˚

90˚

90˚

302˚

150˚165˚

195˚240˚

295˚385˚

315˚

295˚

355˚

315˚

385˚In-Situcoordinates

In-Situcoordinates

270˚ Up,Dn

10 pAm / Division 2

1 pAm / Division 2

E

Tilt-corrected coordinates

105˚NRM

370˚

385˚340˚

Tilt-corrected coordinates

302˚315˚

255˚

355˚

385˚

Kirschvink et al., Fig. S7 E, F

Page 32: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

0

100

200

300

400

500

600

700

800

900

-90 0 270Declination

cm

Section 10at Hijirigawa

Yabe

ina

Zone

R

N

Iwat

o Fo

rmat

ion

trans

iona

l inte

rval

a

c

d

f

g

i

j

m

n

12

34

578

10

1112

10.5

Kirschvink et al.-Fig. S8

Page 33: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

0

10

20

30

-90 270Declination

Stratigraphic coordinates

0

Kirschvink et al.-Fig. S9

m

Alatoconchidae

fusuline

black lime mudstone

dark gray wackestone

R

RN

N

light gray wackestone

rugose coral

?

algal banding

Neo

schw

ager

ina

crat

icul

ifera

Zon

e

Section 2at Saraito

legend

Iwat

o Fo

rmat

ion

Page 34: Palaeogeography, Palaeoclimatology, Palaeoecology...d Dept. Earth & Env. Sci., Kumamoto Univ., Kumamoto 860-8555 Japan e Dept. Earth Planet. Sci., Kobe Univ., Kobe 657-8501, Japan

0

10

20

30

-90 270Declination

Stratigraphic coordinates

0

m

?

R

R

R

N

N

Kirschvink et al.-Fig.S10

Codo

nofu

siella

-Rei

chel

ina

Zone

Lepi

dolin

a Zo

neba

rren

inte

rval

Section 8 at Shioinouso

G-LB fusuline

G-LB chemo

Mita

i For

mat

ion

Iwat

o Fo

rmat

ion

main extinction